Merge pull request #671 from Ana06/release2

Release capa v2.0 🎉

.gitattributes

@@ -0,0 +1,9 @@
+# Set the default behavior, in case people don't have core.autocrlf set.
+* text=auto
+
+# Explicitly declare text files you want to always be normalized and converted
+# to native line endings on checkout.
+*.py text
+*.yml text
+*.md text
+*.txt text

.github/CODE_OF_CONDUCT.md

@@ -1,46 +1,46 @@
-# Contributor Covenant Code of Conduct
-
-## Our Pledge
-
-In the interest of fostering an open and welcoming environment, we as contributors and maintainers pledge to making participation in our project and our community a harassment-free experience for everyone, regardless of age, body size, disability, ethnicity, gender identity and expression, level of experience, nationality, personal appearance, race, religion, or sexual identity and orientation.
-
-## Our Standards
-
-Examples of behavior that contributes to creating a positive environment include:
-
-* Using welcoming and inclusive language
-* Being respectful of differing viewpoints and experiences
-* Gracefully accepting constructive criticism
-* Focusing on what is best for the community
-* Showing empathy towards other community members
-
-Examples of unacceptable behavior by participants include:
-
-* The use of sexualized language or imagery and unwelcome sexual attention or advances
-* Trolling, insulting/derogatory comments, and personal or political attacks
-* Public or private harassment
-* Publishing others' private information, such as a physical or electronic address, without explicit permission
-* Other conduct which could reasonably be considered inappropriate in a professional setting
-
-## Our Responsibilities
-
-Project maintainers are responsible for clarifying the standards of acceptable behavior and are expected to take appropriate and fair corrective action in response to any instances of unacceptable behavior.
-
-Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, or to ban temporarily or permanently any contributor for other behaviors that they deem inappropriate, threatening, offensive, or harmful.
-
-## Scope
-
-This Code of Conduct applies both within project spaces and in public spaces when an individual is representing the project or its community. Examples of representing a project or community include using an official project e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. Representation of a project may be further defined and clarified by project maintainers.
-
-## Enforcement
-
-Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by contacting the project team. All complaints will be reviewed and investigated and will result in a response that is deemed necessary and appropriate to the circumstances. The project team is obligated to maintain confidentiality with regard to the reporter of an incident. Further details of specific enforcement policies may be posted separately.
-
-Project maintainers who do not follow or enforce the Code of Conduct in good faith may face temporary or permanent repercussions as determined by other members of the project's leadership.
-
-## Attribution
-
-This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4, available at [https://contributor-covenant.org/version/1/4][version]
-
-[homepage]: https://contributor-covenant.org
-[version]: https://contributor-covenant.org/version/1/4/
+# Contributor Covenant Code of Conduct
+
+## Our Pledge
+
+In the interest of fostering an open and welcoming environment, we as contributors and maintainers pledge to making participation in our project and our community a harassment-free experience for everyone, regardless of age, body size, disability, ethnicity, gender identity and expression, level of experience, nationality, personal appearance, race, religion, or sexual identity and orientation.
+
+## Our Standards
+
+Examples of behavior that contributes to creating a positive environment include:
+
+* Using welcoming and inclusive language
+* Being respectful of differing viewpoints and experiences
+* Gracefully accepting constructive criticism
+* Focusing on what is best for the community
+* Showing empathy towards other community members
+
+Examples of unacceptable behavior by participants include:
+
+* The use of sexualized language or imagery and unwelcome sexual attention or advances
+* Trolling, insulting/derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others' private information, such as a physical or electronic address, without explicit permission
+* Other conduct which could reasonably be considered inappropriate in a professional setting
+
+## Our Responsibilities
+
+Project maintainers are responsible for clarifying the standards of acceptable behavior and are expected to take appropriate and fair corrective action in response to any instances of unacceptable behavior.
+
+Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, or to ban temporarily or permanently any contributor for other behaviors that they deem inappropriate, threatening, offensive, or harmful.
+
+## Scope
+
+This Code of Conduct applies both within project spaces and in public spaces when an individual is representing the project or its community. Examples of representing a project or community include using an official project e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. Representation of a project may be further defined and clarified by project maintainers.
+
+## Enforcement
+
+Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by contacting the project team. All complaints will be reviewed and investigated and will result in a response that is deemed necessary and appropriate to the circumstances. The project team is obligated to maintain confidentiality with regard to the reporter of an incident. Further details of specific enforcement policies may be posted separately.
+
+Project maintainers who do not follow or enforce the Code of Conduct in good faith may face temporary or permanent repercussions as determined by other members of the project's leadership.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4, available at [https://contributor-covenant.org/version/1/4][version]
+
+[homepage]: https://contributor-covenant.org
+[version]: https://contributor-covenant.org/version/1/4/

.github/CONTRIBUTING.md

@@ -1,197 +1,197 @@
-# Contributing to Capa
-
-First off, thanks for taking the time to contribute!
-
-The following is a set of guidelines for contributing to capa and its packages, which are hosted in the [FireEye Organization](https://github.com/fireeye) on GitHub. These are mostly guidelines, not rules. Use your best judgment, and feel free to propose changes to this document in a pull request.
-
-#### Table Of Contents
-
-[Code of Conduct](#code-of-conduct)
-
-[What should I know before I get started?](#what-should-i-know-before-i-get-started)
-  * [Capa and its Repositories](#capa-and-its-repositories)
-  * [Capa Design Decisions](#design-decisions)
-
-[How Can I Contribute?](#how-can-i-contribute)
-  * [Reporting Bugs](#reporting-bugs)
-  * [Suggesting Enhancements](#suggesting-enhancements)
-  * [Your First Code Contribution](#your-first-code-contribution)
-  * [Pull Requests](#pull-requests)
-
-[Styleguides](#styleguides)
-  * [Git Commit Messages](#git-commit-messages)
-  * [Python Styleguide](#python-styleguide)
-  * [Rules Styleguide](#rules-styleguide)
-
-## Code of Conduct
-
-This project and everyone participating in it is governed by the [Capa Code of Conduct](CODE_OF_CONDUCT.md). By participating, you are expected to uphold this code. Please report unacceptable behavior to the maintainers.
-
-## What should I know before I get started?
-
-### Capa and its repositories
-
-We host the capa project as three Github repositories:
-  - [capa](https://github.com/fireeye/capa)
-  - [capa-rules](https://github.com/fireeye/capa-rules)
-  - [capa-testfiles](https://github.com/fireeye/capa-testfiles)
-  
-The command line tools, logic engine, and other Python source code are found in the `capa` repository.
-This is the repository to fork when you want to enhance the features, performance, or user interface of capa.
-Do *not* push rules directly to this repository, instead...
-
-The standard rules contributed by the community are found in the `capa-rules` repository.
-When you have an idea for a new rule, you should open a PR against `capa-rules`.
-We keep `capa` and `capa-rules` separate to distinguish where ideas, bugs, and discussions should happen.
-If you're writing yaml it probably goes in `capa-rules` and if you're writing Python it probably goes in `capa`.
-Also, we encourage users to develop their own rule repositories, so we treat our default set of rules in the same way.
-
-Test fixtures, such as malware samples and analysis workspaces, are found in the `capa-testfiles` repository.
-These are files you'll need in order to run the linter (in `--thorough` mode) and full test suites;
- however, they take up a lot of space (1GB+), so by keeping `capa-testfiles` separate,
- a shallow checkout of `capa` and `capa-rules` doesn't take much bandwidth.
-
-### Design Decisions
-
-When we make a significant decision in how we maintain the project and what we can or cannot support,
- we will document it in the [capa issues tracker](https://github.com/fireeye/capa/issues).
-This is the best place review our discussions about what/how/why we do things in the project.
-If you have a question, check to see if it is documented there.
-If it is *not* documented there, or you can't find an answer, please open a issue.
-We'll link to existing issues when appropriate to keep discussions in one place.
-
-## How Can I Contribute?
-
-### Reporting Bugs
-
-This section guides you through submitting a bug report for capa.
-Following these guidelines helps maintainers and the community understand your report, reproduce the behavior, and find related reports.
-
-Before creating bug reports, please check [this list](#before-submitting-a-bug-report)
- as you might find out that you don't need to create one.
-When you are creating a bug report, please [include as many details as possible](#how-do-i-submit-a-good-bug-report).
-Fill out [the required template](./ISSUE_TEMPLATE/bug_report.md),
- the information it asks for helps us resolve issues faster.
-
-> **Note:** If you find a **Closed** issue that seems like it is the same thing that you're experiencing, open a new issue and include a link to the original issue in the body of your new one.
-
-#### Before Submitting A Bug Report
-
-* **Determine [which repository the problem should be reported in](#capa-and-its-repositories)**.
-* **Perform a [cursory search](https://github.com/fireeye/capa/issues?q=is%3Aissue)** to see if the problem has already been reported. If it has **and the issue is still open**, add a comment to the existing issue instead of opening a new one.
-
-#### How Do I Submit A (Good) Bug Report?
-
-Bugs are tracked as [GitHub issues](https://guides.github.com/features/issues/).
-After you've determined [which repository](#capa-and-its-repositories) your bug is related to,
- create an issue on that repository and provide the following information by filling in 
- [the template](./ISSUE_TEMPLATE/bug_report.md).
-
-Explain the problem and include additional details to help maintainers reproduce the problem:
-
-* **Use a clear and descriptive title** for the issue to identify the problem.
-* **Describe the exact steps which reproduce the problem** in as many details as possible. For example, start by explaining how you started capa, e.g. which command exactly you used in the terminal, or how you started capa otherwise.
-* **Provide specific examples to demonstrate the steps**. Include links to files or GitHub projects, or copy/pasteable snippets, which you use in those examples. If you're providing snippets in the issue, use [Markdown code blocks](https://help.github.com/articles/markdown-basics/#multiple-lines).
-* **Describe the behavior you observed after following the steps** and point out what exactly is the problem with that behavior.
-* **Explain which behavior you expected to see instead and why.**
-* **Include screenshots and animated GIFs** which show you following the described steps and clearly demonstrate the problem. You can use [this tool](https://www.cockos.com/licecap/) to record GIFs on macOS and Windows, and [this tool](https://github.com/colinkeenan/silentcast) or [this tool](https://github.com/GNOME/byzanz) on Linux.
-* **If you're reporting that capa crashed**, include the stack trace from the terminal. Include the stack trace in the issue in a [code block](https://help.github.com/articles/markdown-basics/#multiple-lines), a [file attachment](https://help.github.com/articles/file-attachments-on-issues-and-pull-requests/), or put it in a [gist](https://gist.github.com/) and provide link to that gist.
-* **If the problem wasn't triggered by a specific action**, describe what you were doing before the problem happened and share more information using the guidelines below.
-
-Provide more context by answering these questions:
-
-* **Did the problem start happening recently** (e.g. after updating to a new version of capa) or was this always a problem?
-* If the problem started happening recently, **can you reproduce the problem in an older version of capa?** What's the most recent version in which the problem doesn't happen? You can download older versions of capa from [the releases page](https://github.com/fireeye/capa/releases).
-* **Can you reliably reproduce the issue?** If not, provide details about how often the problem happens and under which conditions it normally happens.
-* If the problem is related to working with files (e.g. opening and editing files), **does the problem happen for all files and projects or only some?** Does the problem happen only when working with local or remote files (e.g. on network drives), with files of a specific type (e.g. only JavaScript or Python files), with large files or files with very long lines, or with files in a specific encoding? Is there anything else special about the files you are using?
-
-Include details about your configuration and environment:
-
-* **Which version of capa are you using?** You can get the exact version by running `capa --version` in your terminal.
-* **What's the name and version of the OS you're using**?
-
-### Suggesting Enhancements
-
-This section guides you through submitting an enhancement suggestion for capa, including completely new features and minor improvements to existing functionality. Following these guidelines helps maintainers and the community understand your suggestion and find related suggestions.
-
-Before creating enhancement suggestions, please check [this list](#before-submitting-an-enhancement-suggestion) as you might find out that you don't need to create one. When you are creating an enhancement suggestion, please [include as many details as possible](#how-do-i-submit-a-good-enhancement-suggestion). Fill in [the template](./ISSUE_TEMPLATE/feature_request.md), including the steps that you imagine you would take if the feature you're requesting existed.
-
-#### Before Submitting An Enhancement Suggestion
-
-* **Determine [which repository the enhancement should be suggested in](#capa-and-its-repositories).**
-* **Perform a [cursory search](https://github.com/fireeye/capa/issues?q=is%3Aissue)** to see if the enhancement has already been suggested. If it has, add a comment to the existing issue instead of opening a new one.
-
-#### How Do I Submit A (Good) Enhancement Suggestion?
-
-Enhancement suggestions are tracked as [GitHub issues](https://guides.github.com/features/issues/). After you've determined [which repository](#capa-and-its-repositories) your enhancement suggestion is related to, create an issue on that repository and provide the following information:
-
-* **Use a clear and descriptive title** for the issue to identify the suggestion.
-* **Provide a step-by-step description of the suggested enhancement** in as many details as possible.
-* **Provide specific examples to demonstrate the steps**. Include copy/pasteable snippets which you use in those examples, as [Markdown code blocks](https://help.github.com/articles/markdown-basics/#multiple-lines).
-* **Describe the current behavior** and **explain which behavior you expected to see instead** and why.
-* **Include screenshots and animated GIFs** which help you demonstrate the steps or point out the part of capa which the suggestion is related to. You can use [this tool](https://www.cockos.com/licecap/) to record GIFs on macOS and Windows, and [this tool](https://github.com/colinkeenan/silentcast) or [this tool](https://github.com/GNOME/byzanz) on Linux.
-* **Explain why this enhancement would be useful** to most capa users and isn't something that can or should be implemented as an external tool that uses capa as a library.
-* **Specify which version of capa you're using.** You can get the exact version by running `capa --version` in your terminal.
-* **Specify the name and version of the OS you're using.**
-
-### Your First Code Contribution
-
-Unsure where to begin contributing to capa? You can start by looking through these `good-first-issue` and `rule-idea` issues:
-
-* [good-first-issue](https://github.com/fireeye/capa/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) - issues which should only require a few lines of code, and a test or two.
-* [rule-idea](https://github.com/fireeye/capa-rules/issues?q=is%3Aissue+is%3Aopen+label%3A%22rule+idea%22) - issues that describe potential new rule ideas.
-
-Both issue lists are sorted by total number of comments. While not perfect, number of comments is a reasonable proxy for impact a given change will have.
-
-#### Local development
-
-capa and all its resources can be developed locally. 
-For instructions on how to do this, see the "Method 3" section of the [installation guide](https://github.com/fireeye/capa/blob/master/doc/installation.md).
-
-### Pull Requests
-
-The process described here has several goals:
-
-- Maintain capa's quality
-- Fix problems that are important to users
-- Engage the community in working toward the best possible capa
-- Enable a sustainable system for capa's maintainers to review contributions
-
-Please follow these steps to have your contribution considered by the maintainers:
-
-1. Follow all instructions in [the template](PULL_REQUEST_TEMPLATE.md)
-2. Follow the [styleguides](#styleguides)
-3. After you submit your pull request, verify that all [status checks](https://help.github.com/articles/about-status-checks/) are passing <details><summary>What if the status checks are failing? </summary>If a status check is failing, and you believe that the failure is unrelated to your change, please leave a comment on the pull request explaining why you believe the failure is unrelated. A maintainer will re-run the status check for you. If we conclude that the failure was a false positive, then we will open an issue to track that problem with our status check suite.</details>
-
-While the prerequisites above must be satisfied prior to having your pull request reviewed, the reviewer(s) may ask you to complete additional design work, tests, or other changes before your pull request can be ultimately accepted.
-
-## Styleguides
-
-### Git Commit Messages
-
-* Use the present tense ("Add feature" not "Added feature")
-* Use the imperative mood ("Move cursor to..." not "Moves cursor to...")
-* Prefix the first line with the component in question ("rules: ..." or "render: ...")
-* Reference issues and pull requests liberally after the first line
-
-### Python Styleguide
-
-All Python code must adhere to the style guide used by capa: 
-
-  1. [PEP8](https://www.python.org/dev/peps/pep-0008/), with clarifications from
-  2. [Willi's style guide](https://docs.google.com/document/d/1iRpeg-w4DtibwytUyC_dDT7IGhNGBP25-nQfuBa-Fyk/edit?usp=sharing), formatted with
-  3. [isort](https://pypi.org/project/isort/) (with line width 120 and ordered by line length), and formatted with
-  4. [black](https://github.com/psf/black) (with line width 120), and formatted with
-  5. [dos2unix](https://linux.die.net/man/1/dos2unix)
-  
-Our CI pipeline will reformat and enforce the Python styleguide.
- 
-### Rules Styleguide
-
-All (non-nursery) capa rules must:
-
-  1. pass the [linter](https://github.com/fireeye/capa/blob/master/scripts/lint.py), and
-  2. be formatted with [capafmt](https://github.com/fireeye/capa/blob/master/scripts/capafmt.py)
-  
-This ensures that all rules meet the same minimum level of quality and are structured in a consistent way.
-Our CI pipeline will reformat and enforce the capa rules styleguide.
+# Contributing to Capa
+
+First off, thanks for taking the time to contribute!
+
+The following is a set of guidelines for contributing to capa and its packages, which are hosted in the [FireEye Organization](https://github.com/fireeye) on GitHub. These are mostly guidelines, not rules. Use your best judgment, and feel free to propose changes to this document in a pull request.
+
+#### Table Of Contents
+
+[Code of Conduct](#code-of-conduct)
+
+[What should I know before I get started?](#what-should-i-know-before-i-get-started)
+  * [Capa and its Repositories](#capa-and-its-repositories)
+  * [Capa Design Decisions](#design-decisions)
+
+[How Can I Contribute?](#how-can-i-contribute)
+  * [Reporting Bugs](#reporting-bugs)
+  * [Suggesting Enhancements](#suggesting-enhancements)
+  * [Your First Code Contribution](#your-first-code-contribution)
+  * [Pull Requests](#pull-requests)
+
+[Styleguides](#styleguides)
+  * [Git Commit Messages](#git-commit-messages)
+  * [Python Styleguide](#python-styleguide)
+  * [Rules Styleguide](#rules-styleguide)
+
+## Code of Conduct
+
+This project and everyone participating in it is governed by the [Capa Code of Conduct](CODE_OF_CONDUCT.md). By participating, you are expected to uphold this code. Please report unacceptable behavior to the maintainers.
+
+## What should I know before I get started?
+
+### Capa and its repositories
+
+We host the capa project as three Github repositories:
+  - [capa](https://github.com/fireeye/capa)
+  - [capa-rules](https://github.com/fireeye/capa-rules)
+  - [capa-testfiles](https://github.com/fireeye/capa-testfiles)
+  
+The command line tools, logic engine, and other Python source code are found in the `capa` repository.
+This is the repository to fork when you want to enhance the features, performance, or user interface of capa.
+Do *not* push rules directly to this repository, instead...
+
+The standard rules contributed by the community are found in the `capa-rules` repository.
+When you have an idea for a new rule, you should open a PR against `capa-rules`.
+We keep `capa` and `capa-rules` separate to distinguish where ideas, bugs, and discussions should happen.
+If you're writing yaml it probably goes in `capa-rules` and if you're writing Python it probably goes in `capa`.
+Also, we encourage users to develop their own rule repositories, so we treat our default set of rules in the same way.
+
+Test fixtures, such as malware samples and analysis workspaces, are found in the `capa-testfiles` repository.
+These are files you'll need in order to run the linter (in `--thorough` mode) and full test suites;
+ however, they take up a lot of space (1GB+), so by keeping `capa-testfiles` separate,
+ a shallow checkout of `capa` and `capa-rules` doesn't take much bandwidth.
+
+### Design Decisions
+
+When we make a significant decision in how we maintain the project and what we can or cannot support,
+ we will document it in the [capa issues tracker](https://github.com/fireeye/capa/issues).
+This is the best place review our discussions about what/how/why we do things in the project.
+If you have a question, check to see if it is documented there.
+If it is *not* documented there, or you can't find an answer, please open a issue.
+We'll link to existing issues when appropriate to keep discussions in one place.
+
+## How Can I Contribute?
+
+### Reporting Bugs
+
+This section guides you through submitting a bug report for capa.
+Following these guidelines helps maintainers and the community understand your report, reproduce the behavior, and find related reports.
+
+Before creating bug reports, please check [this list](#before-submitting-a-bug-report)
+ as you might find out that you don't need to create one.
+When you are creating a bug report, please [include as many details as possible](#how-do-i-submit-a-good-bug-report).
+Fill out [the required template](./ISSUE_TEMPLATE/bug_report.md),
+ the information it asks for helps us resolve issues faster.
+
+> **Note:** If you find a **Closed** issue that seems like it is the same thing that you're experiencing, open a new issue and include a link to the original issue in the body of your new one.
+
+#### Before Submitting A Bug Report
+
+* **Determine [which repository the problem should be reported in](#capa-and-its-repositories)**.
+* **Perform a [cursory search](https://github.com/fireeye/capa/issues?q=is%3Aissue)** to see if the problem has already been reported. If it has **and the issue is still open**, add a comment to the existing issue instead of opening a new one.
+
+#### How Do I Submit A (Good) Bug Report?
+
+Bugs are tracked as [GitHub issues](https://guides.github.com/features/issues/).
+After you've determined [which repository](#capa-and-its-repositories) your bug is related to,
+ create an issue on that repository and provide the following information by filling in 
+ [the template](./ISSUE_TEMPLATE/bug_report.md).
+
+Explain the problem and include additional details to help maintainers reproduce the problem:
+
+* **Use a clear and descriptive title** for the issue to identify the problem.
+* **Describe the exact steps which reproduce the problem** in as many details as possible. For example, start by explaining how you started capa, e.g. which command exactly you used in the terminal, or how you started capa otherwise.
+* **Provide specific examples to demonstrate the steps**. Include links to files or GitHub projects, or copy/pasteable snippets, which you use in those examples. If you're providing snippets in the issue, use [Markdown code blocks](https://help.github.com/articles/markdown-basics/#multiple-lines).
+* **Describe the behavior you observed after following the steps** and point out what exactly is the problem with that behavior.
+* **Explain which behavior you expected to see instead and why.**
+* **Include screenshots and animated GIFs** which show you following the described steps and clearly demonstrate the problem. You can use [this tool](https://www.cockos.com/licecap/) to record GIFs on macOS and Windows, and [this tool](https://github.com/colinkeenan/silentcast) or [this tool](https://github.com/GNOME/byzanz) on Linux.
+* **If you're reporting that capa crashed**, include the stack trace from the terminal. Include the stack trace in the issue in a [code block](https://help.github.com/articles/markdown-basics/#multiple-lines), a [file attachment](https://help.github.com/articles/file-attachments-on-issues-and-pull-requests/), or put it in a [gist](https://gist.github.com/) and provide link to that gist.
+* **If the problem wasn't triggered by a specific action**, describe what you were doing before the problem happened and share more information using the guidelines below.
+
+Provide more context by answering these questions:
+
+* **Did the problem start happening recently** (e.g. after updating to a new version of capa) or was this always a problem?
+* If the problem started happening recently, **can you reproduce the problem in an older version of capa?** What's the most recent version in which the problem doesn't happen? You can download older versions of capa from [the releases page](https://github.com/fireeye/capa/releases).
+* **Can you reliably reproduce the issue?** If not, provide details about how often the problem happens and under which conditions it normally happens.
+* If the problem is related to working with files (e.g. opening and editing files), **does the problem happen for all files and projects or only some?** Does the problem happen only when working with local or remote files (e.g. on network drives), with files of a specific type (e.g. only JavaScript or Python files), with large files or files with very long lines, or with files in a specific encoding? Is there anything else special about the files you are using?
+
+Include details about your configuration and environment:
+
+* **Which version of capa are you using?** You can get the exact version by running `capa --version` in your terminal.
+* **What's the name and version of the OS you're using**?
+
+### Suggesting Enhancements
+
+This section guides you through submitting an enhancement suggestion for capa, including completely new features and minor improvements to existing functionality. Following these guidelines helps maintainers and the community understand your suggestion and find related suggestions.
+
+Before creating enhancement suggestions, please check [this list](#before-submitting-an-enhancement-suggestion) as you might find out that you don't need to create one. When you are creating an enhancement suggestion, please [include as many details as possible](#how-do-i-submit-a-good-enhancement-suggestion). Fill in [the template](./ISSUE_TEMPLATE/feature_request.md), including the steps that you imagine you would take if the feature you're requesting existed.
+
+#### Before Submitting An Enhancement Suggestion
+
+* **Determine [which repository the enhancement should be suggested in](#capa-and-its-repositories).**
+* **Perform a [cursory search](https://github.com/fireeye/capa/issues?q=is%3Aissue)** to see if the enhancement has already been suggested. If it has, add a comment to the existing issue instead of opening a new one.
+
+#### How Do I Submit A (Good) Enhancement Suggestion?
+
+Enhancement suggestions are tracked as [GitHub issues](https://guides.github.com/features/issues/). After you've determined [which repository](#capa-and-its-repositories) your enhancement suggestion is related to, create an issue on that repository and provide the following information:
+
+* **Use a clear and descriptive title** for the issue to identify the suggestion.
+* **Provide a step-by-step description of the suggested enhancement** in as many details as possible.
+* **Provide specific examples to demonstrate the steps**. Include copy/pasteable snippets which you use in those examples, as [Markdown code blocks](https://help.github.com/articles/markdown-basics/#multiple-lines).
+* **Describe the current behavior** and **explain which behavior you expected to see instead** and why.
+* **Include screenshots and animated GIFs** which help you demonstrate the steps or point out the part of capa which the suggestion is related to. You can use [this tool](https://www.cockos.com/licecap/) to record GIFs on macOS and Windows, and [this tool](https://github.com/colinkeenan/silentcast) or [this tool](https://github.com/GNOME/byzanz) on Linux.
+* **Explain why this enhancement would be useful** to most capa users and isn't something that can or should be implemented as an external tool that uses capa as a library.
+* **Specify which version of capa you're using.** You can get the exact version by running `capa --version` in your terminal.
+* **Specify the name and version of the OS you're using.**
+
+### Your First Code Contribution
+
+Unsure where to begin contributing to capa? You can start by looking through these `good-first-issue` and `rule-idea` issues:
+
+* [good-first-issue](https://github.com/fireeye/capa/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) - issues which should only require a few lines of code, and a test or two.
+* [rule-idea](https://github.com/fireeye/capa-rules/issues?q=is%3Aissue+is%3Aopen+label%3A%22rule+idea%22) - issues that describe potential new rule ideas.
+
+Both issue lists are sorted by total number of comments. While not perfect, number of comments is a reasonable proxy for impact a given change will have.
+
+#### Local development
+
+capa and all its resources can be developed locally. 
+For instructions on how to do this, see the "Method 3" section of the [installation guide](https://github.com/fireeye/capa/blob/master/doc/installation.md).
+
+### Pull Requests
+
+The process described here has several goals:
+
+- Maintain capa's quality
+- Fix problems that are important to users
+- Engage the community in working toward the best possible capa
+- Enable a sustainable system for capa's maintainers to review contributions
+
+Please follow these steps to have your contribution considered by the maintainers:
+
+1. Follow the [styleguides](#styleguides)
+2. Update the CHANGELOG and add tests and documentation. In case they are not needed, indicate it in [the PR template](pull_request_template.md).
+3. After you submit your pull request, verify that all [status checks](https://help.github.com/articles/about-status-checks/) are passing <details><summary>What if the status checks are failing? </summary>If a status check is failing, and you believe that the failure is unrelated to your change, please leave a comment on the pull request explaining why you believe the failure is unrelated. A maintainer will re-run the status check for you. If we conclude that the failure was a false positive, then we will open an issue to track that problem with our status check suite.</details>
+
+While the prerequisites above must be satisfied prior to having your pull request reviewed, the reviewer(s) may ask you to complete additional design work, tests, or other changes before your pull request can be ultimately accepted.
+
+## Styleguides
+
+### Git Commit Messages
+
+* Use the present tense ("Add feature" not "Added feature")
+* Use the imperative mood ("Move cursor to..." not "Moves cursor to...")
+* Prefix the first line with the component in question ("rules: ..." or "render: ...")
+* Reference issues and pull requests liberally after the first line
+
+### Python Styleguide
+
+All Python code must adhere to the style guide used by capa: 
+
+  1. [PEP8](https://www.python.org/dev/peps/pep-0008/), with clarifications from
+  2. [Willi's style guide](https://docs.google.com/document/d/1iRpeg-w4DtibwytUyC_dDT7IGhNGBP25-nQfuBa-Fyk/edit?usp=sharing), formatted with
+  3. [isort](https://pypi.org/project/isort/) (with line width 120 and ordered by line length), and formatted with
+  4. [black](https://github.com/psf/black) (with line width 120), and formatted with
+  5. [dos2unix](https://linux.die.net/man/1/dos2unix)
+  
+Our CI pipeline will reformat and enforce the Python styleguide.
+ 
+### Rules Styleguide
+
+All (non-nursery) capa rules must:
+
+  1. pass the [linter](https://github.com/fireeye/capa/blob/master/scripts/lint.py), and
+  2. be formatted with [capafmt](https://github.com/fireeye/capa/blob/master/scripts/capafmt.py)
+  
+This ensures that all rules meet the same minimum level of quality and are structured in a consistent way.
+Our CI pipeline will reformat and enforce the capa rules styleguide.

.github/ISSUE_TEMPLATE/bug_report.md

@@ -1,47 +1,47 @@
----
-name: Bug report
-about: Create a report to help us improve
-
----
-<!--
-# Is your bug report related to capa rules (for example a false positive)?
-We use sybmodules to separate code, rules and test data. If your issue is related to capa rules, please report it at https://github.com/fireeye/capa-rules/issues.
-
-# Have you checked that your issue isn't already filed?
-Please search if there is a similar issue at https://github.com/fireeye/capa/issues. If there is already a similar issue, please add more details there instead of opening a new one.
-
-# Have you read capa's Code of Conduct?
-By filing an Issue, you are expected to comply with it, including treating everyone with respect: https://github.com/fireeye/capa/blob/master/.github/CODE_OF_CONDUCT.md
-
-# Have you read capa's CONTRIBUTING guide?
-It contains helpful information about how to contribute to capa. Check https://github.com/fireeye/capa/blob/master/.github/CONTRIBUTING.md#reporting-bugs
--->
-
-### Description
-
-<!-- Description of the issue -->
-
-### Steps to Reproduce
-
-<!-- 1. First Step -->
-<!-- 2. Second Step -->
-<!-- 3. and so on… -->
-
-**Expected behavior:**
-
-<!-- What you expect to happen -->
-
-**Actual behavior:**
-
-<!-- What actually happens -->
-
-### Versions
-
-<!-- You can get this information from copy and pasting the output of `capa --version` from the command line.
- Please specify the component you're using (e.g. standalone tool or IDA Pro integration) and your Python version.
- Also, please include the OS and what version of the OS you're running. -->
-
-### Additional Information
-
-<!-- Any additional information, configuration or data that might be necessary to reproduce the issue. -->
-
+---
+name: Bug report
+about: Create a report to help us improve
+
+---
+<!--
+# Is your bug report related to capa rules (for example a false positive)?
+We use submodules to separate code, rules and test data. If your issue is related to capa rules, please report it at https://github.com/fireeye/capa-rules/issues.
+
+# Have you checked that your issue isn't already filed?
+Please search if there is a similar issue at https://github.com/fireeye/capa/issues. If there is already a similar issue, please add more details there instead of opening a new one.
+
+# Have you read capa's Code of Conduct?
+By filing an Issue, you are expected to comply with it, including treating everyone with respect: https://github.com/fireeye/capa/blob/master/.github/CODE_OF_CONDUCT.md
+
+# Have you read capa's CONTRIBUTING guide?
+It contains helpful information about how to contribute to capa. Check https://github.com/fireeye/capa/blob/master/.github/CONTRIBUTING.md#reporting-bugs
+-->
+
+### Description
+
+<!-- Description of the issue -->
+
+### Steps to Reproduce
+
+<!-- 1. First Step -->
+<!-- 2. Second Step -->
+<!-- 3. and so on… -->
+
+**Expected behavior:**
+
+<!-- What you expect to happen -->
+
+**Actual behavior:**
+
+<!-- What actually happens -->
+
+### Versions
+
+<!-- You can get this information from copy and pasting the output of `capa --version` from the command line.
+ Please specify the component you're using (e.g. standalone tool or IDA Pro integration) and your Python version.
+ Also, please include the OS and what version of the OS you're running. -->
+
+### Additional Information
+
+<!-- Any additional information, configuration or data that might be necessary to reproduce the issue. -->
+

.github/ISSUE_TEMPLATE/feature_request.md

@@ -1,35 +1,35 @@
----
-name: Feature request
-about: Suggest an idea for capa
-
----
-<!--
-# Is your issue related to capa rules (for example an idea for a new rule)?
-We use sybmodules to separate code, rules and test data. If your issue is related to capa rules, please report it at https://github.com/fireeye/capa-rules/issues.
-
-# Have you checked that your issue isn't already filed?
-Please search if there is a similar issue at https://github.com/fireeye/capa/issues. If there is already a similar issue, please add more details there instead of opening a new one.
-
-# Have you read capa's Code of Conduct?
-By filing an Issue, you are expected to comply with it, including treating everyone with respect: https://github.com/fireeye/capa/blob/master/.github/CODE_OF_CONDUCT.md
-
-# Have you read capa's CONTRIBUTING guide?
-It contains helpful information about how to contribute to capa. Check https://github.com/fireeye/capa/blob/master/.github/CONTRIBUTING.md#suggesting-enhancements
--->
-
-### Summary
-
-<!-- One paragraph explanation of the feature. -->
-
-### Motivation
-
-<!-- Why are we doing this? What use cases does it support? What is the expected outcome? -->
-
-### Describe alternatives you've considered
-
-<!-- A clear and concise description of the alternative solutions you've considered. -->
-
-## Additional context
-
-<!-- Add any other context or screenshots about the feature request here. -->
-
+---
+name: Feature request
+about: Suggest an idea for capa
+
+---
+<!--
+# Is your issue related to capa rules (for example an idea for a new rule)?
+We use submodules to separate code, rules and test data. If your issue is related to capa rules, please report it at https://github.com/fireeye/capa-rules/issues.
+
+# Have you checked that your issue isn't already filed?
+Please search if there is a similar issue at https://github.com/fireeye/capa/issues. If there is already a similar issue, please add more details there instead of opening a new one.
+
+# Have you read capa's Code of Conduct?
+By filing an Issue, you are expected to comply with it, including treating everyone with respect: https://github.com/fireeye/capa/blob/master/.github/CODE_OF_CONDUCT.md
+
+# Have you read capa's CONTRIBUTING guide?
+It contains helpful information about how to contribute to capa. Check https://github.com/fireeye/capa/blob/master/.github/CONTRIBUTING.md#suggesting-enhancements
+-->
+
+### Summary
+
+<!-- One paragraph explanation of the feature. -->
+
+### Motivation
+
+<!-- Why are we doing this? What use cases does it support? What is the expected outcome? -->
+
+### Describe alternatives you've considered
+
+<!-- A clear and concise description of the alternative solutions you've considered. -->
+
+## Additional context
+
+<!-- Add any other context or screenshots about the feature request here. -->
+

.github/mypy/mypy.ini

@@ -0,0 +1,67 @@
+[mypy]
+
+[mypy-halo.*]
+ignore_missing_imports = True
+
+[mypy-tqdm.*]
+ignore_missing_imports = True
+
+[mypy-ruamel.*]
+ignore_missing_imports = True
+
+[mypy-networkx.*]
+ignore_missing_imports = True
+
+[mypy-pefile.*]
+ignore_missing_imports = True
+
+[mypy-viv_utils.*]
+ignore_missing_imports = True
+
+[mypy-flirt.*]
+ignore_missing_imports = True
+
+[mypy-smda.*]
+ignore_missing_imports = True
+
+[mypy-lief.*]
+ignore_missing_imports = True
+
+[mypy-idc.*]
+ignore_missing_imports = True
+
+[mypy-vivisect.*]
+ignore_missing_imports = True
+
+[mypy-envi.*]
+ignore_missing_imports = True
+
+[mypy-PE.*]
+ignore_missing_imports = True
+
+[mypy-idaapi.*]
+ignore_missing_imports = True
+
+[mypy-idautils.*]
+ignore_missing_imports = True
+
+[mypy-ida_bytes.*]
+ignore_missing_imports = True
+
+[mypy-ida_kernwin.*]
+ignore_missing_imports = True
+
+[mypy-ida_settings.*]
+ignore_missing_imports = True
+
+[mypy-ida_funcs.*]
+ignore_missing_imports = True
+
+[mypy-PyQt5.*]
+ignore_missing_imports = True
+
+[mypy-binaryninja.*]
+ignore_missing_imports = True
+
+[mypy-pytest.*]
+ignore_missing_imports = True

.github/pull_request_template.md

@@ -0,0 +1,22 @@
+
+<!--
+Thank you for contributing to capa! <3
+
+Please read capa's CONTRIBUTING guide if you haven't done so already.
+It contains helpful information about how to contribute to capa. Check https://github.com/fireeye/capa/blob/master/.github/CONTRIBUTING.md
+
+Please describe the changes in this pull request (PR). Include your motivation and context to help us review.
+
+Please mention the issue your PR addresses (if any):
+closes #issue_number
+-->
+
+
+### Checklist
+
+<!-- CHANGELOG.md has a `master (unreleased)` section. Please add bug fixes, new features, breaking changes and anything else you think is worthwhile mentioning in the release notes to this file. -->
+- [ ] No CHANGELOG update needed
+<!-- Tests prove that your fix/work as expected and ensure it doesn't break on the feature. -->
+- [ ] No new tests needed
+<!-- Please help us keeping capa documentation up-to-date -->
+- [ ] No documentation update needed

.github/pyinstaller/pyinstaller.spec

@@ -33,6 +33,7 @@ a = Analysis(
         # this gets invoked from the directory of the spec file,
         # i.e. ./.github/pyinstaller
         ('../../rules', 'rules'),
+        ('../../sigs', 'sigs'),
 
         # capa.render.default uses tabulate that depends on wcwidth.
         # it seems wcwidth uses a json file `version.json`
@@ -108,5 +109,4 @@ exe = EXE(pyz,
 #               a.datas,
 #               strip=None,
 #               upx=True,
-#               name='capa-dat')
-
+#               name='capa-dat')

.github/workflows/build.yml

@@ -1,77 +1,112 @@
-name: build
-
-on:
-  release:
-    types: [edited, published]
-
-jobs:
-  build:
-    name: PyInstaller for ${{ matrix.os }}
-    runs-on: ${{ matrix.os }}
-    strategy:
-      matrix:
-        include:
-          - os: ubuntu-16.04
-            # use old linux so that the shared library versioning is more portable
-            artifact_name: capa
-            asset_name: linux
-          - os: windows-latest
-            artifact_name: capa.exe
-            asset_name: windows
-          - os: macos-latest
-            artifact_name: capa
-            asset_name: macos
-    steps:
-      - name: Checkout capa
-        uses: actions/checkout@v2
-        with:
-          submodules: true
-      - name: Set up Python 3.9
-        uses: actions/setup-python@v2
-        with:
-          python-version: 3.9
-      - if: matrix.os == 'ubuntu-latest'
-        run: sudo apt-get install -y libyaml-dev
-      - name: Install PyInstaller
-        run: pip install 'pyinstaller==4.2'
-      - name: Install capa
-        run: pip install -e .
-      - name: Build standalone executable
-        run: pyinstaller .github/pyinstaller/pyinstaller.spec
-      - name: Does it run?
-        run: dist/capa "tests/data/Practical Malware Analysis Lab 01-01.dll_"
-      - uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.asset_name }}
-          path: dist/${{ matrix.artifact_name }}
-
-  zip:
-    name: zip ${{ matrix.asset_name }}
-    runs-on: ubuntu-latest
-    needs: build
-    strategy:
-      matrix:
-        include:
-          - asset_name: linux
-            artifact_name: capa
-          - asset_name: windows
-            artifact_name: capa.exe
-          - asset_name: macos
-            artifact_name: capa
-    steps:
-      - name: Download ${{ matrix.asset_name }}
-        uses: actions/download-artifact@v2
-        with:
-          name: ${{ matrix.asset_name }}
-      - name: Set executable flag
-        run: chmod +x ${{ matrix.artifact_name }}
-      - name: Set zip name
-        run: echo "zip_name=capa-${GITHUB_REF#refs/tags/}-${{ matrix.asset_name }}.zip" >> $GITHUB_ENV
-      - name: Zip ${{ matrix.artifact_name }} into ${{ env.zip_name }}
-        run: zip ${{ env.zip_name }} ${{ matrix.artifact_name }}
-      - name: Upload ${{ env.zip_name }} to GH Release
-        uses: svenstaro/upload-release-action@v2
-        with:
-          repo_token: ${{ secrets.GITHUB_TOKEN}}
-          file: ${{ env.zip_name }}
-          tag: ${{ github.ref }}
+name: build
+
+on:
+  push:
+    branches: [master]
+  release:
+    types: [edited, published]
+
+jobs:
+  build:
+    name: PyInstaller for ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        include:
+          - os: ubuntu-16.04
+            # use old linux so that the shared library versioning is more portable
+            artifact_name: capa
+            asset_name: linux
+          - os: windows-2019
+            artifact_name: capa.exe
+            asset_name: windows
+          - os: macos-10.15
+            artifact_name: capa
+            asset_name: macos
+    steps:
+      - name: Checkout capa
+        uses: actions/checkout@v2
+        with:
+          submodules: true
+      # using Python 3.8 to support running across multiple operating systems including Windows 7
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.8
+      - if: matrix.os == 'ubuntu-16.04'
+        run: sudo apt-get install -y libyaml-dev
+      - name: Install PyInstaller
+        run: pip install 'pyinstaller==4.2'
+      - name: Install capa
+        run: pip install -e .
+      - name: Build standalone executable
+        run: pyinstaller .github/pyinstaller/pyinstaller.spec
+      - name: Does it run?
+        run: dist/capa "tests/data/Practical Malware Analysis Lab 01-01.dll_"
+      - uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.asset_name }}
+          path: dist/${{ matrix.artifact_name }}
+
+  test_run:
+    # test that binaries run on push to master
+    if: github.event_name == 'push'
+    name: Test run on ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+    needs: [build]
+    strategy:
+      matrix:
+        include:
+          # OSs not already tested above
+          - os: ubuntu-18.04
+            artifact_name: capa
+            asset_name: linux
+          - os: ubuntu-20.04
+            artifact_name: capa
+            asset_name: linux
+          - os: windows-2016
+            artifact_name: capa.exe
+            asset_name: windows
+    steps:
+      - name: Download ${{ matrix.asset_name }}
+        uses: actions/download-artifact@v2
+        with:
+          name: ${{ matrix.asset_name }}
+      - name: Set executable flag
+        if: matrix.os != 'windows-2016'
+        run: chmod +x ${{ matrix.artifact_name }}
+      - name: Run capa
+        run: ./${{ matrix.artifact_name }} -h
+
+  zip_and_upload:
+    # upload zipped binaries to Release page
+    if: github.event_name == 'release'
+    name: zip and upload ${{ matrix.asset_name }}
+    runs-on: ubuntu-20.04
+    needs: [build]
+    strategy:
+      matrix:
+        include:
+          - asset_name: linux
+            artifact_name: capa
+          - asset_name: windows
+            artifact_name: capa.exe
+          - asset_name: macos
+            artifact_name: capa
+    steps:
+      - name: Download ${{ matrix.asset_name }}
+        uses: actions/download-artifact@v2
+        with:
+          name: ${{ matrix.asset_name }}
+      - name: Set executable flag
+        run: chmod +x ${{ matrix.artifact_name }}
+      - name: Set zip name
+        run: echo "zip_name=capa-${GITHUB_REF#refs/tags/}-${{ matrix.asset_name }}.zip" >> $GITHUB_ENV
+      - name: Zip ${{ matrix.artifact_name }} into ${{ env.zip_name }}
+        run: zip ${{ env.zip_name }} ${{ matrix.artifact_name }}
+      - name: Upload ${{ env.zip_name }} to GH Release
+        uses: svenstaro/upload-release-action@v2
+        with:
+          repo_token: ${{ secrets.GITHUB_TOKEN}}
+          file: ${{ env.zip_name }}
+          tag: ${{ github.ref }}

.github/workflows/changelog.yml

@@ -0,0 +1,41 @@
+name: changelog
+
+on:
+  # We need pull_request_target instead of pull_request because a write
+  # repository token is needed to add a review to a PR. DO NOT BUILD
+  # OR RUN UNTRUSTED CODE FROM PRs IN THIS ACTION
+  pull_request_target:
+    types: [opened, edited, synchronize]
+
+jobs:
+  check_changelog:
+    # no need to check for dependency updates via dependabot
+    if: github.actor != 'dependabot[bot]' && github.actor != 'dependabot-preview[bot]'
+    runs-on: ubuntu-20.04
+    env:
+      NO_CHANGELOG: '[x] No CHANGELOG update needed'
+    steps:
+    - name: Get changed files
+      id: files
+      uses: Ana06/get-changed-files@v1.2
+    - name: check changelog updated
+      id: changelog_updated
+      env:
+        PR_BODY: ${{ github.event.pull_request.body }}
+        FILES: ${{ steps.files.outputs.modified }}
+      run: |
+        echo $FILES | grep -qF 'CHANGELOG.md' || echo $PR_BODY | grep -qiF "$NO_CHANGELOG"
+    - name: Reject pull request if no CHANGELOG update
+      if: ${{ always() && steps.changelog_updated.outcome == 'failure' }}
+      uses: Ana06/automatic-pull-request-review@v0.1.0
+      with:
+        repo-token: ${{ secrets.GITHUB_TOKEN }}
+        event: REQUEST_CHANGES
+        body: "Please add bug fixes, new features, breaking changes and anything else you think is worthwhile mentioning to the `master (unreleased)` section of CHANGELOG.md. If no CHANGELOG update is needed add the following to the PR description: `${{ env.NO_CHANGELOG }}`"
+        allow_duplicate: false
+    - name: Dismiss previous review if CHANGELOG update
+      uses: Ana06/automatic-pull-request-review@v0.1.0
+      with:
+        repo-token: ${{ secrets.GITHUB_TOKEN }}
+        event: DISMISS
+        body: "CHANGELOG updated or no update needed, thanks! :smile:"

.github/workflows/publish.yml

@@ -1,29 +1,30 @@
-# This workflows will upload a Python Package using Twine when a release is created
-# For more information see: https://help.github.com/en/actions/language-and-framework-guides/using-python-with-github-actions#publishing-to-package-registries
-
-name: publish to pypi
-
-on:
-  release:
-    types: [published]
-
-jobs:
-  deploy:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python
-        uses: actions/setup-python@v2
-        with:
-          python-version: '2.7'
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          pip install setuptools wheel twine
-      - name: Build and publish
-        env:
-          TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
-          TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
-        run: |
-          python setup.py sdist bdist_wheel
-          twine upload --skip-existing dist/*
+# This workflows will upload a Python Package using Twine when a release is created
+# For more information see: https://help.github.com/en/actions/language-and-framework-guides/using-python-with-github-actions#publishing-to-package-registries
+
+name: publish to pypi
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  deploy:
+    runs-on: ubuntu-20.04
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Python
+        uses: actions/setup-python@v2
+        with:
+          python-version: '3.6'
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install setuptools wheel twine
+      - name: Build and publish
+        env:
+          TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
+          TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
+        run: |
+          python setup.py sdist bdist_wheel
+          twine upload --skip-existing dist/*
+

.github/workflows/tag.yml

@@ -0,0 +1,29 @@
+name: tag
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  tag:
+    name: Tag capa rules
+    runs-on: ubuntu-20.04
+    steps:
+    - name: Checkout capa-rules
+      uses: actions/checkout@v2
+      with:
+        repository: fireeye/capa-rules
+        token: ${{ secrets.CAPA_TOKEN }}
+    - name: Tag capa-rules
+      run: |
+        # user information is needed to create annotated tags (with a message)
+        git config user.email 'capa-dev@fireeye.com'
+        git config user.name 'Capa Bot'
+        name=${{ github.event.release.tag_name }}
+        git tag $name -m "https://github.com/fireeye/capa/releases/$name"
+    - name: Push tag to capa-rules
+      uses: ad-m/github-push-action@master
+      with:
+        repository: fireeye/capa-rules
+        github_token: ${{ secrets.CAPA_TOKEN }}
+        tags: true

.github/workflows/tests.yml

@@ -6,9 +6,24 @@ on:
   pull_request:
     branches: [ master ]
 
+# save workspaces to speed up testing
+env:
+  CAPA_SAVE_WORKSPACE: "True"
+
 jobs:
+  changelog_format:
+    runs-on: ubuntu-20.04
+    steps:
+    - name: Checkout capa
+      uses: actions/checkout@v2
+    # The sync GH action in capa-rules relies on a single '- *$' in the CHANGELOG file
+    - name: Ensure CHANGELOG has '- *$'
+      run: |
+        number=$(grep '\- *$' CHANGELOG.md | wc -l)
+        if [ $number != 1 ]; then exit 1; fi
+
   code_style:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     steps:
     - name: Checkout capa
       uses: actions/checkout@v2
@@ -17,16 +32,18 @@ jobs:
       with:
         python-version: 3.8
     - name: Install dependencies
-      run: pip install 'isort==5.*' black
+      run: pip install -e .[dev]
     - name: Lint with isort
       run: isort --profile black --length-sort --line-width 120 -c .
     - name: Lint with black
       run: black -l 120 --check .
+    - name: Check types with mypy
+      run: mypy --config-file .github/mypy/mypy.ini capa/ scripts/ tests/
 
   rule_linter:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     steps:
-    - name: Checkout capa with rules submodule
+    - name: Checkout capa with submodules
       uses: actions/checkout@v2
       with:
         submodules: true
@@ -34,37 +51,40 @@ jobs:
       uses: actions/setup-python@v2
       with:
         python-version: 3.8
-    # We don't need vivisect, so we can install capa using Python3
     - name: Install capa
       run: pip install -e .
     - name: Run rule linter
       run: python scripts/lint.py rules/
 
   tests:
-    name: Tests in ${{ matrix.python }}
-    runs-on: ubuntu-latest
+    name: Tests in ${{ matrix.python-version }} on ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
     needs: [code_style, rule_linter]
     strategy:
       fail-fast: false
       matrix:
+        os: [ubuntu-20.04, windows-2019, macos-10.15]
+        # across all operating systems
+        python-version: [3.6, 3.9]
         include:
-          - python: 2.7
-          - python: 3.7
-          - python: 3.8
-          - python: 3.9.1
+          # on Ubuntu run these as well
+          - os: ubuntu-20.04
+            python-version: 3.7
+          - os: ubuntu-20.04
+            python-version: 3.8
     steps:
     - name: Checkout capa with submodules
       uses: actions/checkout@v2
       with:
         submodules: true
-    - name: Set up Python ${{ matrix.python }}
+    - name: Set up Python ${{ matrix.python-version }}
       uses: actions/setup-python@v2
       with:
-        python-version: ${{ matrix.python }}
+        python-version: ${{ matrix.python-version }}
     - name: Install pyyaml
+      if: matrix.os == 'ubuntu-20.04'
       run: sudo apt-get install -y libyaml-dev
     - name: Install capa
       run: pip install -e .[dev]
     - name: Run tests
-      run: pytest tests/
-
+      run: pytest -v tests/

README.md

@@ -1,7 +1,10 @@
-![capa](.github/logo.png)
+![capa](https://github.com/fireeye/capa/blob/master/.github/logo.png)
 
+[![PyPI - Python Version](https://img.shields.io/pypi/pyversions/flare-capa)](https://pypi.org/project/flare-capa)
+[![Last release](https://img.shields.io/github/v/release/fireeye/capa)](https://github.com/fireeye/capa/releases)
+[![Number of rules](https://img.shields.io/badge/rules-579-blue.svg)](https://github.com/fireeye/capa-rules)
 [![CI status](https://github.com/fireeye/capa/workflows/CI/badge.svg)](https://github.com/fireeye/capa/actions?query=workflow%3ACI+event%3Apush+branch%3Amaster)
-[![Number of rules](https://img.shields.io/badge/rules-469-blue.svg)](https://github.com/fireeye/capa-rules)
+[![Downloads](https://img.shields.io/github/downloads/fireeye/capa/total)](https://github.com/fireeye/capa/releases)
 [![License](https://img.shields.io/badge/license-Apache--2.0-green.svg)](LICENSE.txt)
 
 capa detects capabilities in executable files.
@@ -62,16 +65,9 @@ $ capa.exe suspicious.exe
 
 Download stable releases of the standalone capa binaries [here](https://github.com/fireeye/capa/releases). You can run the standalone binaries without installation. capa is a command line tool that should be run from the terminal.
 
-<!--
-Alternatively, you can fetch a nightly build of a standalone binary from one of the following links. These are built using the latest development branch.
-- Windows 64bit: TODO
-- Linux: TODO
-- OSX: TODO
--->
+To use capa as a library or integrate with another tool, see [doc/installation.md](https://github.com/fireeye/capa/blob/master/doc/installation.md) for further setup instructions.
 
-To use capa as a library or integrate with another tool, see [doc/installation.md](doc/installation.md) for further setup instructions.
-
-For more information about how to use capa, see [doc/usage.md](doc/usage.md).
+For more information about how to use capa, see [doc/usage.md](https://github.com/fireeye/capa/blob/master/doc/usage.md).
 
 # example
 
@@ -88,7 +84,7 @@ This is useful for at least two reasons:
   - it shows where within the binary an experienced analyst might study with IDA Pro
 
 ```
-λ capa.exe suspicious.exe -vv
+$ capa.exe suspicious.exe -vv
 ...
 execute shell command and capture output
 namespace   c2/shell
@@ -146,18 +142,21 @@ rule:
 The [github.com/fireeye/capa-rules](https://github.com/fireeye/capa-rules) repository contains hundreds of standard library rules that are distributed with capa.
 Please learn to write rules and contribute new entries as you find interesting techniques in malware.
 
-If you use IDA Pro, then you can use the [capa explorer plugin](capa/ida/plugin/).
-capa explorer lets you quickly identify and navigate to interesting areas of a program and help you build new capa rules out of the features extracted directly from your IDB.
+If you use IDA Pro, then you can use the [capa explorer](https://github.com/fireeye/capa/tree/master/capa/ida/plugin) plugin.
+capa explorer helps you identify interesting areas of a program and build new capa rules using features extracted directly from your IDA Pro database.
 
-![capa + IDA Pro integration](doc/img/explorer_expanded.png)
+![capa + IDA Pro integration](https://github.com/fireeye/capa/blob/master/doc/img/explorer_expanded.png)
 
 # further information
 ## capa
-- [doc/installation](doc/installation.md)
-- [doc/usage](doc/usage.md)
-- [doc/limitations](doc/limitations.md)
-- [Contributing Guide](.github/CONTRIBUTING.md)
+- [Installation](https://github.com/fireeye/capa/blob/master/doc/installation.md)
+- [Usage](https://github.com/fireeye/capa/blob/master/doc/usage.md)
+- [Limitations](https://github.com/fireeye/capa/blob/master/doc/limitations.md)
+- [Contributing Guide](https://github.com/fireeye/capa/blob/master/.github/CONTRIBUTING.md)
 
 ## capa rules
 - [capa-rules repository](https://github.com/fireeye/capa-rules)
 - [capa-rules rule format](https://github.com/fireeye/capa-rules/blob/master/doc/format.md)
+
+## capa testfiles
+The [capa-testfiles repository](https://github.com/fireeye/capa-testfiles) contains the data we use to test capa's code and rules

capa/engine.py

@@ -8,11 +8,23 @@
 
 import copy
 import collections
+from typing import Set, Dict, List, Tuple, Union, Mapping
 
-import capa.features
+import capa.rules
+import capa.features.common
+from capa.features.common import Feature
+
+# a collection of features and the locations at which they are found.
+#
+# used throughout matching as the context in which features are searched:
+# to check if a feature exists, do: `Number(0x10) in features`.
+# to collect the locations of a feature, do: `features[Number(0x10)]`
+#
+# aliased here so that the type can be documented and xref'd.
+FeatureSet = Dict[Feature, Set[int]]
 
 
-class Statement(object):
+class Statement:
     """
     superclass for structural nodes, such as and/or/not.
     this exists to provide a default impl for `__str__` and `__repr__`,
@@ -33,7 +45,7 @@ class Statement(object):
     def __repr__(self):
         return str(self)
 
-    def evaluate(self, ctx):
+    def evaluate(self, features: FeatureSet) -> "Result":
         """
         classes that inherit `Statement` must implement `evaluate`
 
@@ -50,7 +62,7 @@ class Statement(object):
             yield self.child
 
         if hasattr(self, "children"):
-            for child in self.children:
+            for child in getattr(self, "children"):
                 yield child
 
     def replace_child(self, existing, new):
@@ -59,12 +71,13 @@ class Statement(object):
                 self.child = new
 
         if hasattr(self, "children"):
-            for i, child in enumerate(self.children):
+            children = getattr(self, "children")
+            for i, child in enumerate(children):
                 if child is existing:
-                    self.children[i] = new
+                    children[i] = new
 
 
-class Result(object):
+class Result:
     """
     represents the results of an evaluation of statements against features.
 
@@ -78,7 +91,7 @@ class Result(object):
     we need this so that we can render the tree of expressions and their results.
     """
 
-    def __init__(self, success, statement, children, locations=None):
+    def __init__(self, success: bool, statement: Union[Statement, Feature], children: List["Result"], locations=None):
         """
         args:
           success (bool)
@@ -199,38 +212,23 @@ class Subscope(Statement):
         raise ValueError("cannot evaluate a subscope directly!")
 
 
-def topologically_order_rules(rules):
-    """
-    order the given rules such that dependencies show up before dependents.
-    this means that as we match rules, we can add features for the matches, and these
-     will be matched by subsequent rules if they follow this order.
-
-    assumes that the rule dependency graph is a DAG.
-    """
-    # we evaluate `rules` multiple times, so if its a generator, realize it into a list.
-    rules = list(rules)
-    namespaces = capa.rules.index_rules_by_namespace(rules)
-    rules = {rule.name: rule for rule in rules}
-    seen = set([])
-    ret = []
-
-    def rec(rule):
-        if rule.name in seen:
-            return
-
-        for dep in rule.get_dependencies(namespaces):
-            rec(rules[dep])
-
-        ret.append(rule)
-        seen.add(rule.name)
-
-    for rule in rules.values():
-        rec(rule)
-
-    return ret
+# mapping from rule name to list of: (location of match, result object)
+#
+# used throughout matching and rendering to collection the results
+#  of statement evaluation and their locations.
+#
+# to check if a rule matched, do: `"TCP client" in matches`.
+# to find where a rule matched, do: `map(first, matches["TCP client"])`
+# to see how a rule matched, do:
+#
+#     for address, match_details in matches["TCP client"]:
+#         inspect(match_details)
+#
+# aliased here so that the type can be documented and xref'd.
+MatchResults = Mapping[str, List[Tuple[int, Result]]]
 
 
-def match(rules, features, va):
+def match(rules: List["capa.rules.Rule"], features: FeatureSet, va: int) -> Tuple[FeatureSet, MatchResults]:
     """
     Args:
       rules (List[capa.rules.Rule]): these must already be ordered topologically by dependency.
@@ -238,11 +236,11 @@ def match(rules, features, va):
       va (int): location of the features
 
     Returns:
-      Tuple[List[capa.features.Feature], Dict[str, Tuple[int, capa.engine.Result]]]: two-tuple with entries:
-        - list of features used for matching (which may be greater than argument, due to rule match features), and
-        - mapping from rule name to (location of match, result object)
+      Tuple[FeatureSet, MatchResults]: two-tuple with entries:
+        - set of features used for matching (which may be greater than argument, due to rule match features), and
+        - mapping from rule name to [(location of match, result object)]
     """
-    results = collections.defaultdict(list)
+    results = collections.defaultdict(list)  # type: MatchResults
 
     # copy features so that we can modify it
     # without affecting the caller (keep this function pure)
@@ -254,12 +252,12 @@ def match(rules, features, va):
         res = rule.evaluate(features)
         if res:
             results[rule.name].append((va, res))
-            features[capa.features.MatchedRule(rule.name)].add(va)
+            features[capa.features.common.MatchedRule(rule.name)].add(va)
 
             namespace = rule.meta.get("namespace")
             if namespace:
                 while namespace:
-                    features[capa.features.MatchedRule(namespace)].add(va)
+                    features[capa.features.common.MatchedRule(namespace)].add(va)
                     namespace, _, _ = namespace.rpartition("/")
 
     return (features, results)

capa/features/__init__.py

@@ -1,222 +0,0 @@
-# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
-# Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at: [package root]/LICENSE.txt
-# Unless required by applicable law or agreed to in writing, software distributed under the License
-#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and limitations under the License.
-
-import re
-import sys
-import codecs
-import logging
-
-import capa.engine
-
-logger = logging.getLogger(__name__)
-MAX_BYTES_FEATURE_SIZE = 0x100
-
-# thunks may be chained so we specify a delta to control the depth to which these chains are explored
-THUNK_CHAIN_DEPTH_DELTA = 5
-
-# identifiers for supported architectures names that tweak a feature
-# for example, offset/x32
-ARCH_X32 = "x32"
-ARCH_X64 = "x64"
-VALID_ARCH = (ARCH_X32, ARCH_X64)
-
-
-def bytes_to_str(b):
-    if sys.version_info[0] >= 3:
-        return str(codecs.encode(b, "hex").decode("utf-8"))
-    else:
-        return codecs.encode(b, "hex")
-
-
-def hex_string(h):
-    """ render hex string e.g. "0a40b1" as "0A 40 B1" """
-    return " ".join(h[i : i + 2] for i in range(0, len(h), 2)).upper()
-
-
-class Feature(object):
-    def __init__(self, value, arch=None, description=None):
-        """
-        Args:
-          value (any): the value of the feature, such as the number or string.
-          arch (str): one of the VALID_ARCH values, or None.
-            When None, then the feature applies to any architecture.
-            Modifies the feature name from `feature` to `feature/arch`, like `offset/x32`.
-          description (str): a human-readable description that explains the feature value.
-        """
-        super(Feature, self).__init__()
-
-        if arch is not None:
-            if arch not in VALID_ARCH:
-                raise ValueError("arch '%s' must be one of %s" % (arch, VALID_ARCH))
-            self.name = self.__class__.__name__.lower() + "/" + arch
-        else:
-            self.name = self.__class__.__name__.lower()
-
-        self.value = value
-        self.arch = arch
-        self.description = description
-
-    def __hash__(self):
-        return hash((self.name, self.value, self.arch))
-
-    def __eq__(self, other):
-        return self.name == other.name and self.value == other.value and self.arch == other.arch
-
-    def get_value_str(self):
-        """
-        render the value of this feature, for use by `__str__` and friends.
-        subclasses should override to customize the rendering.
-
-        Returns: any
-        """
-        return self.value
-
-    def __str__(self):
-        if self.value is not None:
-            if self.description:
-                return "%s(%s = %s)" % (self.name, self.get_value_str(), self.description)
-            else:
-                return "%s(%s)" % (self.name, self.get_value_str())
-        else:
-            return "%s" % self.name
-
-    def __repr__(self):
-        return str(self)
-
-    def evaluate(self, ctx):
-        return capa.engine.Result(self in ctx, self, [], locations=ctx.get(self, []))
-
-    def freeze_serialize(self):
-        if self.arch is not None:
-            return (self.__class__.__name__, [self.value, {"arch": self.arch}])
-        else:
-            return (self.__class__.__name__, [self.value])
-
-    @classmethod
-    def freeze_deserialize(cls, args):
-        # as you can see below in code,
-        # if the last argument is a dictionary,
-        # consider it to be kwargs passed to the feature constructor.
-        if len(args) == 1:
-            return cls(*args)
-        elif isinstance(args[-1], dict):
-            kwargs = args[-1]
-            args = args[:-1]
-            return cls(*args, **kwargs)
-
-
-class MatchedRule(Feature):
-    def __init__(self, value, description=None):
-        super(MatchedRule, self).__init__(value, description=description)
-        self.name = "match"
-
-
-class Characteristic(Feature):
-    def __init__(self, value, description=None):
-        super(Characteristic, self).__init__(value, description=description)
-
-
-class String(Feature):
-    def __init__(self, value, description=None):
-        super(String, self).__init__(value, description=description)
-
-
-class Regex(String):
-    def __init__(self, value, description=None):
-        super(Regex, self).__init__(value, description=description)
-        pat = self.value[len("/") : -len("/")]
-        flags = re.DOTALL
-        if value.endswith("/i"):
-            pat = self.value[len("/") : -len("/i")]
-            flags |= re.IGNORECASE
-        try:
-            self.re = re.compile(pat, flags)
-        except re.error:
-            if value.endswith("/i"):
-                value = value[: -len("i")]
-            raise ValueError(
-                "invalid regular expression: %s it should use Python syntax, try it at https://pythex.org" % value
-            )
-
-    def evaluate(self, ctx):
-        for feature, locations in ctx.items():
-            if not isinstance(feature, (capa.features.String,)):
-                continue
-
-            # `re.search` finds a match anywhere in the given string
-            # which implies leading and/or trailing whitespace.
-            # using this mode cleans is more convenient for rule authors,
-            # so that they don't have to prefix/suffix their terms like: /.*foo.*/.
-            if self.re.search(feature.value):
-                # unlike other features, we cannot return put a reference to `self` directly in a `Result`.
-                # this is because `self` may match on many strings, so we can't stuff the matched value into it.
-                # instead, return a new instance that has a reference to both the regex and the matched value.
-                # see #262.
-                return capa.engine.Result(True, _MatchedRegex(self, feature.value), [], locations=locations)
-
-        return capa.engine.Result(False, _MatchedRegex(self, None), [])
-
-    def __str__(self):
-        return "regex(string =~ %s)" % self.value
-
-
-class _MatchedRegex(Regex):
-    """
-    this represents a specific instance of a regular expression feature match.
-    treat it the same as a `Regex` except it has the `match` field that contains the complete string that matched.
-
-    note: this type should only ever be constructed by `Regex.evaluate()`. it is not part of the public API.
-    """
-
-    def __init__(self, regex, match):
-        """
-        args:
-          regex (Regex): the regex feature that matches
-          match (string|None): the matching string or None if it doesn't match
-        """
-        super(_MatchedRegex, self).__init__(regex.value, description=regex.description)
-        # we want this to collide with the name of `Regex` above,
-        # so that it works nicely with the renderers.
-        self.name = "regex"
-        # this may be None if the regex doesn't match
-        self.match = match
-
-    def __str__(self):
-        return 'regex(string =~ %s, matched = "%s")' % (self.value, self.match)
-
-
-class StringFactory(object):
-    def __new__(self, value, description=None):
-        if value.startswith("/") and (value.endswith("/") or value.endswith("/i")):
-            return Regex(value, description=description)
-        return String(value, description=description)
-
-
-class Bytes(Feature):
-    def __init__(self, value, description=None):
-        super(Bytes, self).__init__(value, description=description)
-
-    def evaluate(self, ctx):
-        for feature, locations in ctx.items():
-            if not isinstance(feature, (capa.features.Bytes,)):
-                continue
-
-            if feature.value.startswith(self.value):
-                return capa.engine.Result(True, self, [], locations=locations)
-
-        return capa.engine.Result(False, self, [])
-
-    def get_value_str(self):
-        return hex_string(bytes_to_str(self.value))
-
-    def freeze_serialize(self):
-        return (self.__class__.__name__, [bytes_to_str(self.value).upper()])
-
-    @classmethod
-    def freeze_deserialize(cls, args):
-        return cls(*[codecs.decode(x, "hex") for x in args])

capa/features/basicblock.py

@@ -6,7 +6,7 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-from capa.features import Feature
+from capa.features.common import Feature
 
 
 class BasicBlock(Feature):

capa/features/common.py

@@ -0,0 +1,262 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+
+import re
+import codecs
+import logging
+import collections
+from typing import Set, Dict, Union
+
+import capa.engine
+import capa.features
+
+logger = logging.getLogger(__name__)
+MAX_BYTES_FEATURE_SIZE = 0x100
+
+# thunks may be chained so we specify a delta to control the depth to which these chains are explored
+THUNK_CHAIN_DEPTH_DELTA = 5
+
+# identifiers for supported architectures names that tweak a feature
+# for example, offset/x32
+ARCH_X32 = "x32"
+ARCH_X64 = "x64"
+VALID_ARCH = (ARCH_X32, ARCH_X64)
+
+
+def bytes_to_str(b: bytes) -> str:
+    return str(codecs.encode(b, "hex").decode("utf-8"))
+
+
+def hex_string(h: str) -> str:
+    """render hex string e.g. "0a40b1" as "0A 40 B1" """
+    return " ".join(h[i : i + 2] for i in range(0, len(h), 2)).upper()
+
+
+def escape_string(s: str) -> str:
+    """escape special characters"""
+    s = repr(s)
+    if not s.startswith(('"', "'")):
+        # u'hello\r\nworld' -> hello\\r\\nworld
+        s = s[2:-1]
+    else:
+        # 'hello\r\nworld' -> hello\\r\\nworld
+        s = s[1:-1]
+    s = s.replace("\\'", "'")  # repr() may escape "'" in some edge cases, remove
+    s = s.replace('"', '\\"')  # repr() does not escape '"', add
+    return s
+
+
+class Feature:
+    def __init__(self, value: Union[str, int, bytes], arch=None, description=None):
+        """
+        Args:
+          value (any): the value of the feature, such as the number or string.
+          arch (str): one of the VALID_ARCH values, or None.
+            When None, then the feature applies to any architecture.
+            Modifies the feature name from `feature` to `feature/arch`, like `offset/x32`.
+          description (str): a human-readable description that explains the feature value.
+        """
+        super(Feature, self).__init__()
+
+        if arch is not None:
+            if arch not in VALID_ARCH:
+                raise ValueError("arch '%s' must be one of %s" % (arch, VALID_ARCH))
+            self.name = self.__class__.__name__.lower() + "/" + arch
+        else:
+            self.name = self.__class__.__name__.lower()
+
+        self.value = value
+        self.arch = arch
+        self.description = description
+
+    def __hash__(self):
+        return hash((self.name, self.value, self.arch))
+
+    def __eq__(self, other):
+        return self.name == other.name and self.value == other.value and self.arch == other.arch
+
+    def get_value_str(self) -> str:
+        """
+        render the value of this feature, for use by `__str__` and friends.
+        subclasses should override to customize the rendering.
+
+        Returns: any
+        """
+        return str(self.value)
+
+    def __str__(self):
+        if self.value is not None:
+            if self.description:
+                return "%s(%s = %s)" % (self.name, self.get_value_str(), self.description)
+            else:
+                return "%s(%s)" % (self.name, self.get_value_str())
+        else:
+            return "%s" % self.name
+
+    def __repr__(self):
+        return str(self)
+
+    def evaluate(self, ctx: Dict["Feature", Set[int]]) -> "capa.engine.Result":
+        return capa.engine.Result(self in ctx, self, [], locations=ctx.get(self, []))
+
+    def freeze_serialize(self):
+        if self.arch is not None:
+            return (self.__class__.__name__, [self.value, {"arch": self.arch}])
+        else:
+            return (self.__class__.__name__, [self.value])
+
+    @classmethod
+    def freeze_deserialize(cls, args):
+        # as you can see below in code,
+        # if the last argument is a dictionary,
+        # consider it to be kwargs passed to the feature constructor.
+        if len(args) == 1:
+            return cls(*args)
+        elif isinstance(args[-1], dict):
+            kwargs = args[-1]
+            args = args[:-1]
+            return cls(*args, **kwargs)
+
+
+class MatchedRule(Feature):
+    def __init__(self, value: str, description=None):
+        super(MatchedRule, self).__init__(value, description=description)
+        self.name = "match"
+
+
+class Characteristic(Feature):
+    def __init__(self, value: str, description=None):
+        super(Characteristic, self).__init__(value, description=description)
+
+
+class String(Feature):
+    def __init__(self, value: str, description=None):
+        super(String, self).__init__(value, description=description)
+
+
+class Regex(String):
+    def __init__(self, value: str, description=None):
+        super(Regex, self).__init__(value, description=description)
+        self.value = value
+
+        pat = self.value[len("/") : -len("/")]
+        flags = re.DOTALL
+        if value.endswith("/i"):
+            pat = self.value[len("/") : -len("/i")]
+            flags |= re.IGNORECASE
+        try:
+            self.re = re.compile(pat, flags)
+        except re.error:
+            if value.endswith("/i"):
+                value = value[: -len("i")]
+            raise ValueError(
+                "invalid regular expression: %s it should use Python syntax, try it at https://pythex.org" % value
+            )
+
+    def evaluate(self, ctx):
+        # mapping from string value to list of locations.
+        # will unique the locations later on.
+        matches = collections.defaultdict(list)
+
+        for feature, locations in ctx.items():
+            if not isinstance(feature, (String,)):
+                continue
+
+            if not isinstance(feature.value, str):
+                # this is a programming error: String should only contain str
+                raise ValueError("unexpected feature value type")
+
+            # `re.search` finds a match anywhere in the given string
+            # which implies leading and/or trailing whitespace.
+            # using this mode cleans is more convenient for rule authors,
+            # so that they don't have to prefix/suffix their terms like: /.*foo.*/.
+            if self.re.search(feature.value):
+                matches[feature.value].extend(locations)
+
+        if matches:
+            # finalize: defaultdict -> dict
+            # which makes json serialization easier
+            matches = dict(matches)
+
+            # collect all locations
+            locations = set()
+            for s in matches.keys():
+                matches[s] = list(set(matches[s]))
+                locations.update(matches[s])
+
+            # unlike other features, we cannot return put a reference to `self` directly in a `Result`.
+            # this is because `self` may match on many strings, so we can't stuff the matched value into it.
+            # instead, return a new instance that has a reference to both the regex and the matched values.
+            # see #262.
+            return capa.engine.Result(True, _MatchedRegex(self, matches), [], locations=locations)
+        else:
+            return capa.engine.Result(False, _MatchedRegex(self, None), [])
+
+    def __str__(self):
+        return "regex(string =~ %s)" % self.value
+
+
+class _MatchedRegex(Regex):
+    """
+    this represents specific match instances of a regular expression feature.
+    treat it the same as a `Regex` except it has the `matches` field that contains the complete strings that matched.
+
+    note: this type should only ever be constructed by `Regex.evaluate()`. it is not part of the public API.
+    """
+
+    def __init__(self, regex: Regex, matches):
+        """
+        args:
+          regex (Regex): the regex feature that matches.
+          match (Dict[string, List[int]]|None): mapping from matching string to its locations.
+        """
+        super(_MatchedRegex, self).__init__(str(regex.value), description=regex.description)
+        # we want this to collide with the name of `Regex` above,
+        # so that it works nicely with the renderers.
+        self.name = "regex"
+        # this may be None if the regex doesn't match
+        self.matches = matches
+
+    def __str__(self):
+        return "regex(string =~ %s, matches = %s)" % (
+            self.value,
+            ", ".join(map(lambda s: '"' + s + '"', (self.matches or {}).keys())),
+        )
+
+
+class StringFactory:
+    def __new__(cls, value: str, description=None):
+        if value.startswith("/") and (value.endswith("/") or value.endswith("/i")):
+            return Regex(value, description=description)
+        return String(value, description=description)
+
+
+class Bytes(Feature):
+    def __init__(self, value: bytes, description=None):
+        super(Bytes, self).__init__(value, description=description)
+        self.value = value
+
+    def evaluate(self, ctx):
+        for feature, locations in ctx.items():
+            if not isinstance(feature, (Bytes,)):
+                continue
+
+            if feature.value.startswith(self.value):
+                return capa.engine.Result(True, self, [], locations=locations)
+
+        return capa.engine.Result(False, self, [])
+
+    def get_value_str(self):
+        return hex_string(bytes_to_str(self.value))
+
+    def freeze_serialize(self):
+        return (self.__class__.__name__, [bytes_to_str(self.value).upper()])
+
+    @classmethod
+    def freeze_deserialize(cls, args):
+        return cls(*[codecs.decode(x, "hex") for x in args])

capa/features/extractors/__init__.py

@@ -1,286 +0,0 @@
-# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
-# Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at: [package root]/LICENSE.txt
-# Unless required by applicable law or agreed to in writing, software distributed under the License
-#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and limitations under the License.
-
-import abc
-
-
-class FeatureExtractor(object):
-    """
-    FeatureExtractor defines the interface for fetching features from a sample.
-
-    There may be multiple backends that support fetching features for capa.
-    For example, we use vivisect by default, but also want to support saving
-     and restoring features from a JSON file.
-    When we restore the features, we'd like to use exactly the same matching logic
-     to find matching rules.
-    Therefore, we can define a FeatureExtractor that provides features from the
-     serialized JSON file and do matching without a binary analysis pass.
-    Also, this provides a way to hook in an IDA backend.
-
-    This class is not instantiated directly; it is the base class for other implementations.
-    """
-
-    __metaclass__ = abc.ABCMeta
-
-    def __init__(self):
-        #
-        # note: a subclass should define ctor parameters for its own use.
-        #  for example, the Vivisect feature extract might require the vw and/or path.
-        # this base class doesn't know what to do with that info, though.
-        #
-        super(FeatureExtractor, self).__init__()
-
-    @abc.abstractmethod
-    def get_base_address(self):
-        """
-        fetch the preferred load address at which the sample was analyzed.
-
-        returns: int
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def extract_file_features(self):
-        """
-        extract file-scope features.
-
-        example::
-
-            extractor = VivisectFeatureExtractor(vw, path)
-            for feature, va in extractor.get_file_features():
-                print('0x%x: %s', va, feature)
-
-        yields:
-          Tuple[capa.features.Feature, int]: feature and its location
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def get_functions(self):
-        """
-        enumerate the functions and provide opaque values that will
-         subsequently be provided to `.extract_function_features()`, etc.
-
-        by "opaque value", we mean that this can be any object, as long as it
-         provides enough context to `.extract_function_features()`.
-
-        the opaque value should support casting to int (`__int__`) for the function start address.
-
-        yields:
-          any: the opaque function value.
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def extract_function_features(self, f):
-        """
-        extract function-scope features.
-        the arguments are opaque values previously provided by `.get_functions()`, etc.
-
-        example::
-
-            extractor = VivisectFeatureExtractor(vw, path)
-            for function in extractor.get_functions():
-                for feature, va in extractor.extract_function_features(function):
-                    print('0x%x: %s', va, feature)
-
-        args:
-          f [any]: an opaque value previously fetched from `.get_functions()`.
-
-        yields:
-          Tuple[capa.features.Feature, int]: feature and its location
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def get_basic_blocks(self, f):
-        """
-        enumerate the basic blocks in the given function and provide opaque values that will
-         subsequently be provided to `.extract_basic_block_features()`, etc.
-
-        by "opaque value", we mean that this can be any object, as long as it
-         provides enough context to `.extract_basic_block_features()`.
-
-        the opaque value should support casting to int (`__int__`) for the basic block start address.
-
-        yields:
-          any: the opaque basic block value.
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def extract_basic_block_features(self, f, bb):
-        """
-        extract basic block-scope features.
-        the arguments are opaque values previously provided by `.get_functions()`, etc.
-
-        example::
-
-            extractor = VivisectFeatureExtractor(vw, path)
-            for function in extractor.get_functions():
-                for bb in extractor.get_basic_blocks(function):
-                    for feature, va in extractor.extract_basic_block_features(function, bb):
-                        print('0x%x: %s', va, feature)
-
-        args:
-          f [any]: an opaque value previously fetched from `.get_functions()`.
-          bb [any]: an opaque value previously fetched from `.get_basic_blocks()`.
-
-        yields:
-          Tuple[capa.features.Feature, int]: feature and its location
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def get_instructions(self, f, bb):
-        """
-        enumerate the instructions in the given basic block and provide opaque values that will
-         subsequently be provided to `.extract_insn_features()`, etc.
-
-        by "opaque value", we mean that this can be any object, as long as it
-         provides enough context to `.extract_insn_features()`.
-
-        the opaque value should support casting to int (`__int__`) for the instruction address.
-
-        yields:
-          any: the opaque function value.
-        """
-        raise NotImplemented
-
-    @abc.abstractmethod
-    def extract_insn_features(self, f, bb, insn):
-        """
-        extract instruction-scope features.
-        the arguments are opaque values previously provided by `.get_functions()`, etc.
-
-        example::
-
-            extractor = VivisectFeatureExtractor(vw, path)
-            for function in extractor.get_functions():
-                for bb in extractor.get_basic_blocks(function):
-                    for insn in extractor.get_instructions(function, bb):
-                        for feature, va in extractor.extract_insn_features(function, bb, insn):
-                            print('0x%x: %s', va, feature)
-
-        args:
-          f [any]: an opaque value previously fetched from `.get_functions()`.
-          bb [any]: an opaque value previously fetched from `.get_basic_blocks()`.
-          insn [any]: an opaque value previously fetched from `.get_instructions()`.
-
-        yields:
-          Tuple[capa.features.Feature, int]: feature and its location
-        """
-        raise NotImplemented
-
-
-class NullFeatureExtractor(FeatureExtractor):
-    """
-    An extractor that extracts some user-provided features.
-    The structure of the single parameter is demonstrated in the example below.
-
-    This is useful for testing, as we can provide expected values and see if matching works.
-    Also, this is how we represent features deserialized from a freeze file.
-
-    example::
-
-        extractor = NullFeatureExtractor({
-            'base address: 0x401000,
-            'file features': [
-                (0x402345, capa.features.Characteristic('embedded pe')),
-            ],
-            'functions': {
-                0x401000: {
-                    'features': [
-                        (0x401000, capa.features.Characteristic('nzxor')),
-                    ],
-                    'basic blocks': {
-                        0x401000: {
-                            'features': [
-                                (0x401000, capa.features.Characteristic('tight-loop')),
-                            ],
-                            'instructions': {
-                                0x401000: {
-                                    'features': [
-                                        (0x401000, capa.features.Characteristic('nzxor')),
-                                    ],
-                                },
-                                0x401002: ...
-                            }
-                        },
-                        0x401005: ...
-                    }
-                },
-                0x40200: ...
-            }
-        )
-    """
-
-    def __init__(self, features):
-        super(NullFeatureExtractor, self).__init__()
-        self.features = features
-
-    def get_base_address(self):
-        return self.features["base address"]
-
-    def extract_file_features(self):
-        for p in self.features.get("file features", []):
-            va, feature = p
-            yield feature, va
-
-    def get_functions(self):
-        for va in sorted(self.features["functions"].keys()):
-            yield va
-
-    def extract_function_features(self, f):
-        for p in self.features.get("functions", {}).get(f, {}).get("features", []):  # noqa: E127 line over-indented
-            va, feature = p
-            yield feature, va
-
-    def get_basic_blocks(self, f):
-        for va in sorted(
-            self.features.get("functions", {})  # noqa: E127 line over-indented
-            .get(f, {})
-            .get("basic blocks", {})
-            .keys()
-        ):
-            yield va
-
-    def extract_basic_block_features(self, f, bb):
-        for p in (
-            self.features.get("functions", {})  # noqa: E127 line over-indented
-            .get(f, {})
-            .get("basic blocks", {})
-            .get(bb, {})
-            .get("features", [])
-        ):
-            va, feature = p
-            yield feature, va
-
-    def get_instructions(self, f, bb):
-        for va in sorted(
-            self.features.get("functions", {})  # noqa: E127 line over-indented
-            .get(f, {})
-            .get("basic blocks", {})
-            .get(bb, {})
-            .get("instructions", {})
-            .keys()
-        ):
-            yield va
-
-    def extract_insn_features(self, f, bb, insn):
-        for p in (
-            self.features.get("functions", {})  # noqa: E127 line over-indented
-            .get(f, {})
-            .get("basic blocks", {})
-            .get(bb, {})
-            .get("instructions", {})
-            .get(insn, {})
-            .get("features", [])
-        ):
-            va, feature = p
-            yield feature, va

capa/features/extractors/base_extractor.py

@@ -0,0 +1,312 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+
+import abc
+from typing import Tuple, Iterator, SupportsInt
+
+from capa.features.basicblock import Feature
+
+# feature extractors may reference functions, BBs, insns by opaque handle values.
+# the only requirement of these handles are that they support `__int__`,
+# so that they can be rendered as addresses.
+#
+# these handles are only consumed by routines on
+# the feature extractor from which they were created.
+#
+# int(FunctionHandle) -> function start address
+# int(BBHandle) -> BasicBlock start address
+# int(InsnHandle) -> instruction address
+FunctionHandle = SupportsInt
+BBHandle = SupportsInt
+InsnHandle = SupportsInt
+
+
+class FeatureExtractor:
+    """
+    FeatureExtractor defines the interface for fetching features from a sample.
+
+    There may be multiple backends that support fetching features for capa.
+    For example, we use vivisect by default, but also want to support saving
+     and restoring features from a JSON file.
+    When we restore the features, we'd like to use exactly the same matching logic
+     to find matching rules.
+    Therefore, we can define a FeatureExtractor that provides features from the
+     serialized JSON file and do matching without a binary analysis pass.
+    Also, this provides a way to hook in an IDA backend.
+
+    This class is not instantiated directly; it is the base class for other implementations.
+    """
+
+    __metaclass__ = abc.ABCMeta
+
+    def __init__(self):
+        #
+        # note: a subclass should define ctor parameters for its own use.
+        #  for example, the Vivisect feature extract might require the vw and/or path.
+        # this base class doesn't know what to do with that info, though.
+        #
+        super(FeatureExtractor, self).__init__()
+
+    @abc.abstractmethod
+    def get_base_address(self) -> int:
+        """
+        fetch the preferred load address at which the sample was analyzed.
+        """
+        raise NotImplemented
+
+    @abc.abstractmethod
+    def extract_file_features(self) -> Iterator[Tuple[Feature, int]]:
+        """
+        extract file-scope features.
+
+        example::
+
+            extractor = VivisectFeatureExtractor(vw, path)
+            for feature, va in extractor.get_file_features():
+                print('0x%x: %s', va, feature)
+
+        yields:
+          Tuple[Feature, int]: feature and its location
+        """
+        raise NotImplemented
+
+    @abc.abstractmethod
+    def get_functions(self) -> Iterator[FunctionHandle]:
+        """
+        enumerate the functions and provide opaque values that will
+         subsequently be provided to `.extract_function_features()`, etc.
+        """
+        raise NotImplemented
+
+    def is_library_function(self, va: int) -> bool:
+        """
+        is the given address a library function?
+        the backend may implement its own function matching algorithm, or none at all.
+        we accept a VA here, rather than function object, to handle addresses identified in instructions.
+
+        this information is used to:
+          - filter out matches in library functions (by default), and
+          - recognize when to fetch symbol names for called (non-API) functions
+
+        args:
+          va (int): the virtual address of a function.
+
+        returns:
+          bool: True if the given address is the start of a library function.
+        """
+        return False
+
+    def get_function_name(self, va: int) -> str:
+        """
+        fetch any recognized name for the given address.
+        this is only guaranteed to return a value when the given function is a recognized library function.
+        we accept a VA here, rather than function object, to handle addresses identified in instructions.
+
+        args:
+          va (int): the virtual address of a function.
+
+        returns:
+          str: the function name
+
+        raises:
+          KeyError: when the given function does not have a name.
+        """
+        raise KeyError(va)
+
+    @abc.abstractmethod
+    def extract_function_features(self, f: FunctionHandle) -> Iterator[Tuple[Feature, int]]:
+        """
+        extract function-scope features.
+        the arguments are opaque values previously provided by `.get_functions()`, etc.
+
+        example::
+
+            extractor = VivisectFeatureExtractor(vw, path)
+            for function in extractor.get_functions():
+                for feature, va in extractor.extract_function_features(function):
+                    print('0x%x: %s', va, feature)
+
+        args:
+          f [FunctionHandle]: an opaque value previously fetched from `.get_functions()`.
+
+        yields:
+          Tuple[Feature, int]: feature and its location
+        """
+        raise NotImplemented
+
+    @abc.abstractmethod
+    def get_basic_blocks(self, f: FunctionHandle) -> Iterator[BBHandle]:
+        """
+        enumerate the basic blocks in the given function and provide opaque values that will
+         subsequently be provided to `.extract_basic_block_features()`, etc.
+        """
+        raise NotImplemented
+
+    @abc.abstractmethod
+    def extract_basic_block_features(self, f: FunctionHandle, bb: BBHandle) -> Iterator[Tuple[Feature, int]]:
+        """
+        extract basic block-scope features.
+        the arguments are opaque values previously provided by `.get_functions()`, etc.
+
+        example::
+
+            extractor = VivisectFeatureExtractor(vw, path)
+            for function in extractor.get_functions():
+                for bb in extractor.get_basic_blocks(function):
+                    for feature, va in extractor.extract_basic_block_features(function, bb):
+                        print('0x%x: %s', va, feature)
+
+        args:
+          f [FunctionHandle]: an opaque value previously fetched from `.get_functions()`.
+          bb [BBHandle]: an opaque value previously fetched from `.get_basic_blocks()`.
+
+        yields:
+          Tuple[Feature, int]: feature and its location
+        """
+        raise NotImplemented
+
+    @abc.abstractmethod
+    def get_instructions(self, f: FunctionHandle, bb: BBHandle) -> Iterator[InsnHandle]:
+        """
+        enumerate the instructions in the given basic block and provide opaque values that will
+         subsequently be provided to `.extract_insn_features()`, etc.
+        """
+        raise NotImplemented
+
+    @abc.abstractmethod
+    def extract_insn_features(self, f: FunctionHandle, bb: BBHandle, insn: InsnHandle) -> Iterator[Tuple[Feature, int]]:
+        """
+        extract instruction-scope features.
+        the arguments are opaque values previously provided by `.get_functions()`, etc.
+
+        example::
+
+            extractor = VivisectFeatureExtractor(vw, path)
+            for function in extractor.get_functions():
+                for bb in extractor.get_basic_blocks(function):
+                    for insn in extractor.get_instructions(function, bb):
+                        for feature, va in extractor.extract_insn_features(function, bb, insn):
+                            print('0x%x: %s', va, feature)
+
+        args:
+          f [FunctionHandle]: an opaque value previously fetched from `.get_functions()`.
+          bb [BBHandle]: an opaque value previously fetched from `.get_basic_blocks()`.
+          insn [InsnHandle]: an opaque value previously fetched from `.get_instructions()`.
+
+        yields:
+          Tuple[Feature, int]: feature and its location
+        """
+        raise NotImplemented
+
+
+class NullFeatureExtractor(FeatureExtractor):
+    """
+    An extractor that extracts some user-provided features.
+    The structure of the single parameter is demonstrated in the example below.
+
+    This is useful for testing, as we can provide expected values and see if matching works.
+    Also, this is how we represent features deserialized from a freeze file.
+
+    example::
+
+        extractor = NullFeatureExtractor({
+            'base address: 0x401000,
+            'file features': [
+                (0x402345, capa.features.Characteristic('embedded pe')),
+            ],
+            'functions': {
+                0x401000: {
+                    'features': [
+                        (0x401000, capa.features.Characteristic('nzxor')),
+                    ],
+                    'basic blocks': {
+                        0x401000: {
+                            'features': [
+                                (0x401000, capa.features.Characteristic('tight-loop')),
+                            ],
+                            'instructions': {
+                                0x401000: {
+                                    'features': [
+                                        (0x401000, capa.features.Characteristic('nzxor')),
+                                    ],
+                                },
+                                0x401002: ...
+                            }
+                        },
+                        0x401005: ...
+                    }
+                },
+                0x40200: ...
+            }
+        )
+    """
+
+    def __init__(self, features):
+        super(NullFeatureExtractor, self).__init__()
+        self.features = features
+
+    def get_base_address(self):
+        return self.features["base address"]
+
+    def extract_file_features(self):
+        for p in self.features.get("file features", []):
+            va, feature = p
+            yield feature, va
+
+    def get_functions(self):
+        for va in sorted(self.features["functions"].keys()):
+            yield va
+
+    def extract_function_features(self, f):
+        for p in self.features.get("functions", {}).get(f, {}).get("features", []):  # noqa: E127 line over-indented
+            va, feature = p
+            yield feature, va
+
+    def get_basic_blocks(self, f):
+        for va in sorted(
+            self.features.get("functions", {})  # noqa: E127 line over-indented
+            .get(f, {})
+            .get("basic blocks", {})
+            .keys()
+        ):
+            yield va
+
+    def extract_basic_block_features(self, f, bb):
+        for p in (
+            self.features.get("functions", {})  # noqa: E127 line over-indented
+            .get(f, {})
+            .get("basic blocks", {})
+            .get(bb, {})
+            .get("features", [])
+        ):
+            va, feature = p
+            yield feature, va
+
+    def get_instructions(self, f, bb):
+        for va in sorted(
+            self.features.get("functions", {})  # noqa: E127 line over-indented
+            .get(f, {})
+            .get("basic blocks", {})
+            .get(bb, {})
+            .get("instructions", {})
+            .keys()
+        ):
+            yield va
+
+    def extract_insn_features(self, f, bb, insn):
+        for p in (
+            self.features.get("functions", {})  # noqa: E127 line over-indented
+            .get(f, {})
+            .get("basic blocks", {})
+            .get(bb, {})
+            .get("instructions", {})
+            .get(insn, {})
+            .get("features", [])
+        ):
+            va, feature = p
+            yield feature, va

capa/features/extractors/helpers.py

@@ -6,23 +6,18 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import sys
+import struct
 import builtins
-
-from capa.features.file import Import
-from capa.features.insn import API
+from typing import Tuple, Iterator
 
 MIN_STACKSTRING_LEN = 8
 
 
-def xor_static(data, i):
-    if sys.version_info >= (3, 0):
-        return bytes(c ^ i for c in data)
-    else:
-        return "".join(chr(ord(c) ^ i) for c in data)
+def xor_static(data: bytes, i: int) -> bytes:
+    return bytes(c ^ i for c in data)
 
 
-def is_aw_function(symbol):
+def is_aw_function(symbol: str) -> bool:
     """
     is the given function name an A/W function?
     these are variants of functions that, on Windows, accept either a narrow or wide string.
@@ -38,7 +33,7 @@ def is_aw_function(symbol):
     return "a" <= symbol[-2] <= "z" or "0" <= symbol[-2] <= "9"
 
 
-def is_ordinal(symbol):
+def is_ordinal(symbol: str) -> bool:
     """
     is the given symbol an ordinal that is prefixed by "#"?
     """
@@ -47,7 +42,7 @@ def is_ordinal(symbol):
     return False
 
 
-def generate_symbols(dll, symbol):
+def generate_symbols(dll: str, symbol: str) -> Iterator[str]:
     """
     for a given dll and symbol name, generate variants.
     we over-generate features to make matching easier.
@@ -73,11 +68,11 @@ def generate_symbols(dll, symbol):
             yield symbol[:-1]
 
 
-def all_zeros(bytez):
+def all_zeros(bytez: bytes) -> bool:
     return all(b == 0 for b in builtins.bytes(bytez))
 
 
-def twos_complement(val, bits):
+def twos_complement(val: int, bits: int) -> int:
     """
     compute the 2's complement of int value val
 
@@ -90,3 +85,49 @@ def twos_complement(val, bits):
     else:
         # return positive value as is
         return val
+
+
+def carve_pe(pbytes: bytes, offset: int = 0) -> Iterator[Tuple[int, int]]:
+    """
+    Generate (offset, key) tuples of embedded PEs
+
+    Based on the version from vivisect:
+      https://github.com/vivisect/vivisect/blob/7be4037b1cecc4551b397f840405a1fc606f9b53/PE/carve.py#L19
+    And its IDA adaptation:
+      capa/features/extractors/ida/file.py
+    """
+    mz_xor = [
+        (
+            xor_static(b"MZ", key),
+            xor_static(b"PE", key),
+            key,
+        )
+        for key in range(256)
+    ]
+
+    pblen = len(pbytes)
+    todo = [(pbytes.find(mzx, offset), mzx, pex, key) for mzx, pex, key in mz_xor]
+    todo = [(off, mzx, pex, key) for (off, mzx, pex, key) in todo if off != -1]
+
+    while len(todo):
+
+        off, mzx, pex, key = todo.pop()
+
+        # The MZ header has one field we will check
+        # e_lfanew is at 0x3c
+        e_lfanew = off + 0x3C
+        if pblen < (e_lfanew + 4):
+            continue
+
+        newoff = struct.unpack("<I", xor_static(pbytes[e_lfanew : e_lfanew + 4], key))[0]
+
+        nextres = pbytes.find(mzx, off + 1)
+        if nextres != -1:
+            todo.append((nextres, mzx, pex, key))
+
+        peoff = off + newoff
+        if pblen < (peoff + 2):
+            continue
+
+        if pbytes[peoff : peoff + 2] == pex:
+            yield (off, key)

capa/features/extractors/ida/__init__.py

@@ -1,93 +0,0 @@
-# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
-# Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at: [package root]/LICENSE.txt
-# Unless required by applicable law or agreed to in writing, software distributed under the License
-#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and limitations under the License.
-
-import sys
-import types
-
-import idaapi
-
-import capa.features.extractors.ida.file
-import capa.features.extractors.ida.insn
-import capa.features.extractors.ida.function
-import capa.features.extractors.ida.basicblock
-from capa.features.extractors import FeatureExtractor
-
-
-def get_ea(self):
-    """ """
-    if isinstance(self, (idaapi.BasicBlock, idaapi.func_t)):
-        return self.start_ea
-    if isinstance(self, idaapi.insn_t):
-        return self.ea
-    raise TypeError
-
-
-def add_ea_int_cast(o):
-    """
-    dynamically add a cast-to-int (`__int__`) method to the given object
-    that returns the value of the `.ea` property.
-    this bit of skullduggery lets use cast viv-utils objects as ints.
-    the correct way of doing this is to update viv-utils (or subclass the objects here).
-    """
-    if sys.version_info[0] >= 3:
-        setattr(o, "__int__", types.MethodType(get_ea, o))
-    else:
-        setattr(o, "__int__", types.MethodType(get_ea, o, type(o)))
-    return o
-
-
-class IdaFeatureExtractor(FeatureExtractor):
-    def __init__(self):
-        super(IdaFeatureExtractor, self).__init__()
-
-    def get_base_address(self):
-        return idaapi.get_imagebase()
-
-    def extract_file_features(self):
-        for (feature, ea) in capa.features.extractors.ida.file.extract_features():
-            yield feature, ea
-
-    def get_functions(self):
-        import capa.features.extractors.ida.helpers as ida_helpers
-
-        # data structure shared across functions yielded here.
-        # useful for caching analysis relevant across a single workspace.
-        ctx = {}
-
-        # ignore library functions and thunk functions as identified by IDA
-        for f in ida_helpers.get_functions(skip_thunks=True, skip_libs=True):
-            setattr(f, "ctx", ctx)
-            yield add_ea_int_cast(f)
-
-    @staticmethod
-    def get_function(ea):
-        f = idaapi.get_func(ea)
-        setattr(f, "ctx", {})
-        return add_ea_int_cast(f)
-
-    def extract_function_features(self, f):
-        for (feature, ea) in capa.features.extractors.ida.function.extract_features(f):
-            yield feature, ea
-
-    def get_basic_blocks(self, f):
-        for bb in capa.features.extractors.ida.helpers.get_function_blocks(f):
-            yield add_ea_int_cast(bb)
-
-    def extract_basic_block_features(self, f, bb):
-        for (feature, ea) in capa.features.extractors.ida.basicblock.extract_features(f, bb):
-            yield feature, ea
-
-    def get_instructions(self, f, bb):
-        import capa.features.extractors.ida.helpers as ida_helpers
-
-        for insn in ida_helpers.get_instructions_in_range(bb.start_ea, bb.end_ea):
-            yield add_ea_int_cast(insn)
-
-    def extract_insn_features(self, f, bb, insn):
-        for (feature, ea) in capa.features.extractors.ida.insn.extract_features(f, bb, insn):
-            yield feature, ea

capa/features/extractors/ida/basicblock.py

@@ -6,14 +6,13 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import sys
 import string
 import struct
 
 import idaapi
 
 import capa.features.extractors.ida.helpers
-from capa.features import Characteristic
+from capa.features.common import Characteristic
 from capa.features.basicblock import BasicBlock
 from capa.features.extractors.ida import helpers
 from capa.features.extractors.helpers import MIN_STACKSTRING_LEN
@@ -39,18 +38,11 @@ def get_printable_len(op):
         raise ValueError("Unhandled operand data type 0x%x." % op.dtype)
 
     def is_printable_ascii(chars):
-        if sys.version_info[0] >= 3:
-            return all(c < 127 and chr(c) in string.printable for c in chars)
-        else:
-            return all(ord(c) < 127 and c in string.printable for c in chars)
+        return all(c < 127 and chr(c) in string.printable for c in chars)
 
     def is_printable_utf16le(chars):
-        if sys.version_info[0] >= 3:
-            if all(c == 0x00 for c in chars[1::2]):
-                return is_printable_ascii(chars[::2])
-        else:
-            if all(c == "\x00" for c in chars[1::2]):
-                return is_printable_ascii(chars[::2])
+        if all(c == 0x00 for c in chars[1::2]):
+            return is_printable_ascii(chars[::2])
 
     if is_printable_ascii(chars):
         return idaapi.get_dtype_size(op.dtype)

capa/features/extractors/ida/extractor.py

@@ -0,0 +1,108 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+
+import idaapi
+
+import capa.features.extractors.ida.file
+import capa.features.extractors.ida.insn
+import capa.features.extractors.ida.function
+import capa.features.extractors.ida.basicblock
+from capa.features.extractors.base_extractor import FeatureExtractor
+
+
+class FunctionHandle:
+    """this acts like an idaapi.func_t but with __int__()"""
+
+    def __init__(self, inner):
+        self._inner = inner
+
+    def __int__(self):
+        return self.start_ea
+
+    def __getattr__(self, name):
+        return getattr(self._inner, name)
+
+
+class BasicBlockHandle:
+    """this acts like an idaapi.BasicBlock but with __int__()"""
+
+    def __init__(self, inner):
+        self._inner = inner
+
+    def __int__(self):
+        return self.start_ea
+
+    def __getattr__(self, name):
+        return getattr(self._inner, name)
+
+
+class InstructionHandle:
+    """this acts like an idaapi.insn_t but with __int__()"""
+
+    def __init__(self, inner):
+        self._inner = inner
+
+    def __int__(self):
+        return self.ea
+
+    def __getattr__(self, name):
+        return getattr(self._inner, name)
+
+
+class IdaFeatureExtractor(FeatureExtractor):
+    def __init__(self):
+        super(IdaFeatureExtractor, self).__init__()
+
+    def get_base_address(self):
+        return idaapi.get_imagebase()
+
+    def extract_file_features(self):
+        for (feature, ea) in capa.features.extractors.ida.file.extract_features():
+            yield feature, ea
+
+    def get_functions(self):
+        import capa.features.extractors.ida.helpers as ida_helpers
+
+        # data structure shared across functions yielded here.
+        # useful for caching analysis relevant across a single workspace.
+        ctx = {}
+
+        # ignore library functions and thunk functions as identified by IDA
+        for f in ida_helpers.get_functions(skip_thunks=True, skip_libs=True):
+            setattr(f, "ctx", ctx)
+            yield FunctionHandle(f)
+
+    @staticmethod
+    def get_function(ea):
+        f = idaapi.get_func(ea)
+        setattr(f, "ctx", {})
+        return FunctionHandle(f)
+
+    def extract_function_features(self, f):
+        for (feature, ea) in capa.features.extractors.ida.function.extract_features(f):
+            yield feature, ea
+
+    def get_basic_blocks(self, f):
+        import capa.features.extractors.ida.helpers as ida_helpers
+
+        for bb in ida_helpers.get_function_blocks(f):
+            yield BasicBlockHandle(bb)
+
+    def extract_basic_block_features(self, f, bb):
+        for (feature, ea) in capa.features.extractors.ida.basicblock.extract_features(f, bb):
+            yield feature, ea
+
+    def get_instructions(self, f, bb):
+        import capa.features.extractors.ida.helpers as ida_helpers
+
+        for insn in ida_helpers.get_instructions_in_range(bb.start_ea, bb.end_ea):
+            yield InstructionHandle(insn)
+
+    def extract_insn_features(self, f, bb, insn):
+        for (feature, ea) in capa.features.extractors.ida.insn.extract_features(f, bb, insn):
+            yield feature, ea

capa/features/extractors/ida/file.py

@@ -15,8 +15,8 @@ import idautils
 import capa.features.extractors.helpers
 import capa.features.extractors.strings
 import capa.features.extractors.ida.helpers
-from capa.features import String, Characteristic
-from capa.features.file import Export, Import, Section
+from capa.features.file import Export, Import, Section, FunctionName
+from capa.features.common import String, Characteristic
 
 
 def check_segment_for_pe(seg):
@@ -78,7 +78,7 @@ def extract_file_embedded_pe():
 
 
 def extract_file_export_names():
-    """ extract function exports """
+    """extract function exports"""
     for (_, _, ea, name) in idautils.Entries():
         yield Export(name), ea
 
@@ -143,8 +143,18 @@ def extract_file_strings():
             yield String(s.s), (seg.start_ea + s.offset)
 
 
+def extract_file_function_names():
+    """
+    extract the names of statically-linked library functions.
+    """
+    for ea in idautils.Functions():
+        if idaapi.get_func(ea).flags & idaapi.FUNC_LIB:
+            name = idaapi.get_name(ea)
+            yield FunctionName(name), ea
+
+
 def extract_features():
-    """ extract file features """
+    """extract file features"""
     for file_handler in FILE_HANDLERS:
         for feature, va in file_handler():
             yield feature, va
@@ -156,6 +166,7 @@ FILE_HANDLERS = (
     extract_file_strings,
     extract_file_section_names,
     extract_file_embedded_pe,
+    extract_file_function_names,
 )
 
 

capa/features/extractors/ida/function.py

@@ -10,7 +10,7 @@ import idaapi
 import idautils
 
 import capa.features.extractors.ida.helpers
-from capa.features import Characteristic
+from capa.features.common import Characteristic
 from capa.features.extractors import loops
 
 

capa/features/extractors/ida/helpers.py

@@ -6,9 +6,6 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import sys
-import string
-
 import idc
 import idaapi
 import idautils
@@ -23,11 +20,7 @@ def find_byte_sequence(start, end, seq):
         end: max virtual address
         seq: bytes to search e.g. b"\x01\x03"
     """
-    if sys.version_info[0] >= 3:
-        seq = " ".join(["%02x" % b for b in seq])
-    else:
-        seq = " ".join(["%02x" % ord(b) for b in seq])
-
+    seq = " ".join(["%02x" % b for b in seq])
     while True:
         ea = idaapi.find_binary(start, end, seq, 0, idaapi.SEARCH_DOWN)
         if ea == idaapi.BADADDR:
@@ -83,7 +76,7 @@ def get_segment_buffer(seg):
 
 
 def get_file_imports():
-    """ get file imports """
+    """get file imports"""
     imports = {}
 
     for idx in range(idaapi.get_import_module_qty()):
@@ -120,7 +113,7 @@ def get_instructions_in_range(start, end):
 
 
 def is_operand_equal(op1, op2):
-    """ compare two IDA op_t """
+    """compare two IDA op_t"""
     if op1.flags != op2.flags:
         return False
 
@@ -146,7 +139,7 @@ def is_operand_equal(op1, op2):
 
 
 def is_basic_block_equal(bb1, bb2):
-    """ compare two IDA BasicBlock """
+    """compare two IDA BasicBlock"""
     if bb1.start_ea != bb2.start_ea:
         return False
 
@@ -160,7 +153,7 @@ def is_basic_block_equal(bb1, bb2):
 
 
 def basic_block_size(bb):
-    """ calculate size of basic block """
+    """calculate size of basic block"""
     return bb.end_ea - bb.start_ea
 
 
@@ -178,7 +171,7 @@ def read_bytes_at(ea, count):
 
 
 def find_string_at(ea, min=4):
-    """ check if ASCII string exists at a given virtual address """
+    """check if ASCII string exists at a given virtual address"""
     found = idaapi.get_strlit_contents(ea, -1, idaapi.STRTYPE_C)
     if found and len(found) > min:
         try:
@@ -232,23 +225,23 @@ def get_op_phrase_info(op):
 
 
 def is_op_write(insn, op):
-    """ Check if an operand is written to (destination operand) """
+    """Check if an operand is written to (destination operand)"""
     return idaapi.has_cf_chg(insn.get_canon_feature(), op.n)
 
 
 def is_op_read(insn, op):
-    """ Check if an operand is read from (source operand) """
+    """Check if an operand is read from (source operand)"""
     return idaapi.has_cf_use(insn.get_canon_feature(), op.n)
 
 
 def is_op_offset(insn, op):
-    """ Check is an operand has been marked as an offset (by auto-analysis or manually) """
+    """Check is an operand has been marked as an offset (by auto-analysis or manually)"""
     flags = idaapi.get_flags(insn.ea)
     return ida_bytes.is_off(flags, op.n)
 
 
 def is_sp_modified(insn):
-    """ determine if instruction modifies SP, ESP, RSP """
+    """determine if instruction modifies SP, ESP, RSP"""
     for op in get_insn_ops(insn, target_ops=(idaapi.o_reg,)):
         if op.reg == idautils.procregs.sp.reg and is_op_write(insn, op):
             # register is stack and written
@@ -257,7 +250,7 @@ def is_sp_modified(insn):
 
 
 def is_bp_modified(insn):
-    """ check if instruction modifies BP, EBP, RBP """
+    """check if instruction modifies BP, EBP, RBP"""
     for op in get_insn_ops(insn, target_ops=(idaapi.o_reg,)):
         if op.reg == idautils.procregs.bp.reg and is_op_write(insn, op):
             # register is base and written
@@ -266,12 +259,12 @@ def is_bp_modified(insn):
 
 
 def is_frame_register(reg):
-    """ check if register is sp or bp """
+    """check if register is sp or bp"""
     return reg in (idautils.procregs.sp.reg, idautils.procregs.bp.reg)
 
 
 def get_insn_ops(insn, target_ops=()):
-    """ yield op_t for instruction, filter on type if specified """
+    """yield op_t for instruction, filter on type if specified"""
     for op in insn.ops:
         if op.type == idaapi.o_void:
             # avoid looping all 6 ops if only subset exists
@@ -282,7 +275,7 @@ def get_insn_ops(insn, target_ops=()):
 
 
 def is_op_stack_var(ea, index):
-    """ check if operand is a stack variable """
+    """check if operand is a stack variable"""
     return idaapi.is_stkvar(idaapi.get_flags(ea), index)
 
 
@@ -336,7 +329,7 @@ def is_basic_block_tight_loop(bb):
 
 
 def find_data_reference_from_insn(insn, max_depth=10):
-    """ search for data reference from instruction, return address of instruction if no reference exists """
+    """search for data reference from instruction, return address of instruction if no reference exists"""
     depth = 0
     ea = insn.ea
 
@@ -379,5 +372,5 @@ def get_function_blocks(f):
 
 
 def is_basic_block_return(bb):
-    """ check if basic block is return block """
+    """check if basic block is return block"""
     return bb.type == idaapi.fcb_ret

capa/features/extractors/ida/insn.py

@@ -12,7 +12,8 @@ import idautils
 
 import capa.features.extractors.helpers
 import capa.features.extractors.ida.helpers
-from capa.features import (
+from capa.features.insn import API, Number, Offset, Mnemonic
+from capa.features.common import (
     ARCH_X32,
     ARCH_X64,
     MAX_BYTES_FEATURE_SIZE,
@@ -21,7 +22,6 @@ from capa.features import (
     String,
     Characteristic,
 )
-from capa.features.insn import API, Number, Offset, Mnemonic
 
 # security cookie checks may perform non-zeroing XORs, these are expected within a certain
 # byte range within the first and returning basic blocks, this helps to reduce FP features
@@ -53,10 +53,7 @@ def get_imports(ctx):
 
 
 def check_for_api_call(ctx, insn):
-    """ check instruction for API call """
-    if not insn.get_canon_mnem() in ("call", "jmp"):
-        return
-
+    """check instruction for API call"""
     info = ()
     ref = insn.ea
 
@@ -95,11 +92,29 @@ def extract_insn_api_features(f, bb, insn):
     example:
         call dword [0x00473038]
     """
+    if not insn.get_canon_mnem() in ("call", "jmp"):
+        return
+
     for api in check_for_api_call(f.ctx, insn):
         dll, _, symbol = api.rpartition(".")
         for name in capa.features.extractors.helpers.generate_symbols(dll, symbol):
             yield API(name), insn.ea
 
+    # extract IDA/FLIRT recognized API functions
+    targets = tuple(idautils.CodeRefsFrom(insn.ea, False))
+    if not targets:
+        return
+
+    target = targets[0]
+    target_func = idaapi.get_func(target)
+    if not target_func or target_func.start_ea != target:
+        # not a function (start)
+        return
+
+    if target_func.flags & idaapi.FUNC_LIB:
+        name = idaapi.get_name(target_func.start_ea)
+        yield API(name), insn.ea
+
 
 def extract_insn_number_features(f, bb, insn):
     """parse instruction number features
@@ -256,7 +271,7 @@ def bb_stack_cookie_registers(bb):
 
 
 def is_nzxor_stack_cookie_delta(f, bb, insn):
-    """ check if nzxor exists within stack cookie delta """
+    """check if nzxor exists within stack cookie delta"""
     # security cookie check should use SP or BP
     if not capa.features.extractors.ida.helpers.is_frame_register(insn.Op2.reg):
         return False
@@ -279,7 +294,7 @@ def is_nzxor_stack_cookie_delta(f, bb, insn):
 
 
 def is_nzxor_stack_cookie(f, bb, insn):
-    """ check if nzxor is related to stack cookie """
+    """check if nzxor is related to stack cookie"""
     if contains_stack_cookie_keywords(idaapi.get_cmt(insn.ea, False)):
         # Example:
         #   xor     ecx, ebp        ; StackCookie

capa/features/extractors/pefile.py

@@ -0,0 +1,173 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+import logging
+
+import pefile
+
+import capa.features.extractors
+import capa.features.extractors.helpers
+import capa.features.extractors.strings
+from capa.features.file import Export, Import, Section
+from capa.features.common import String, Characteristic
+from capa.features.extractors.base_extractor import FeatureExtractor
+
+logger = logging.getLogger(__name__)
+
+
+def extract_file_embedded_pe(pe, file_path):
+    with open(file_path, "rb") as f:
+        fbytes = f.read()
+
+    for offset, i in capa.features.extractors.helpers.carve_pe(fbytes, 1):
+        yield Characteristic("embedded pe"), offset
+
+
+def extract_file_export_names(pe, file_path):
+    base_address = pe.OPTIONAL_HEADER.ImageBase
+
+    if hasattr(pe, "DIRECTORY_ENTRY_EXPORT"):
+        for export in pe.DIRECTORY_ENTRY_EXPORT.symbols:
+            try:
+                name = export.name.partition(b"\x00")[0].decode("ascii")
+            except UnicodeDecodeError:
+                continue
+            va = base_address + export.address
+            yield Export(name), va
+
+
+def extract_file_import_names(pe, file_path):
+    """
+    extract imported function names
+    1. imports by ordinal:
+     - modulename.#ordinal
+    2. imports by name, results in two features to support importname-only matching:
+     - modulename.importname
+     - importname
+    """
+    if hasattr(pe, "DIRECTORY_ENTRY_IMPORT"):
+        for dll in pe.DIRECTORY_ENTRY_IMPORT:
+            try:
+                modname = dll.dll.partition(b"\x00")[0].decode("ascii")
+            except UnicodeDecodeError:
+                continue
+
+            # strip extension
+            modname = modname.rpartition(".")[0].lower()
+
+            for imp in dll.imports:
+                if imp.import_by_ordinal:
+                    impname = "#%s" % imp.ordinal
+                else:
+                    try:
+                        impname = imp.name.partition(b"\x00")[0].decode("ascii")
+                    except UnicodeDecodeError:
+                        continue
+
+                for name in capa.features.extractors.helpers.generate_symbols(modname, impname):
+                    yield Import(name), imp.address
+
+
+def extract_file_section_names(pe, file_path):
+    base_address = pe.OPTIONAL_HEADER.ImageBase
+
+    for section in pe.sections:
+        try:
+            name = section.Name.partition(b"\x00")[0].decode("ascii")
+        except UnicodeDecodeError:
+            continue
+
+        yield Section(name), base_address + section.VirtualAddress
+
+
+def extract_file_strings(pe, file_path):
+    """
+    extract ASCII and UTF-16 LE strings from file
+    """
+    with open(file_path, "rb") as f:
+        b = f.read()
+
+    for s in capa.features.extractors.strings.extract_ascii_strings(b):
+        yield String(s.s), s.offset
+
+    for s in capa.features.extractors.strings.extract_unicode_strings(b):
+        yield String(s.s), s.offset
+
+
+def extract_file_function_names(pe, file_path):
+    """
+    extract the names of statically-linked library functions.
+    """
+    if False:
+        # using a `yield` here to force this to be a generator, not function.
+        yield NotImplementedError("pefile doesn't have library matching")
+    return
+
+
+def extract_file_features(pe, file_path):
+    """
+    extract file features from given workspace
+
+    args:
+      pe (pefile.PE): the parsed PE
+      file_path: path to the input file
+
+    yields:
+      Tuple[Feature, VA]: a feature and its location.
+    """
+
+    for file_handler in FILE_HANDLERS:
+        for feature, va in file_handler(pe, file_path):
+            yield feature, va
+
+
+FILE_HANDLERS = (
+    extract_file_embedded_pe,
+    extract_file_export_names,
+    extract_file_import_names,
+    extract_file_section_names,
+    extract_file_strings,
+    extract_file_function_names,
+)
+
+
+class PefileFeatureExtractor(FeatureExtractor):
+    def __init__(self, path: str):
+        super(PefileFeatureExtractor, self).__init__()
+        self.path = path
+        self.pe = pefile.PE(path)
+
+    def get_base_address(self):
+        return self.pe.OPTIONAL_HEADER.ImageBase
+
+    def extract_file_features(self):
+        for feature, va in extract_file_features(self.pe, self.path):
+            yield feature, va
+
+    def get_functions(self):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def extract_function_features(self, f):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def get_basic_blocks(self, f):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def extract_basic_block_features(self, f, bb):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def get_instructions(self, f, bb):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def extract_insn_features(self, f, bb, insn):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def is_library_function(self, va):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")
+
+    def get_function_name(self, va):
+        raise NotImplementedError("PefileFeatureExtract can only be used to extract file features")

capa/features/extractors/smda/__init__.py

@@ -1,52 +0,0 @@
-import sys
-import types
-
-from smda.common.SmdaReport import SmdaReport
-from smda.common.SmdaInstruction import SmdaInstruction
-
-import capa.features.extractors.smda.file
-import capa.features.extractors.smda.insn
-import capa.features.extractors.smda.function
-import capa.features.extractors.smda.basicblock
-from capa.main import UnsupportedRuntimeError
-from capa.features.extractors import FeatureExtractor
-
-
-class SmdaFeatureExtractor(FeatureExtractor):
-    def __init__(self, smda_report: SmdaReport, path):
-        super(SmdaFeatureExtractor, self).__init__()
-        if sys.version_info < (3, 0):
-            raise UnsupportedRuntimeError("SMDA should only be used with Python 3.")
-        self.smda_report = smda_report
-        self.path = path
-
-    def get_base_address(self):
-        return self.smda_report.base_addr
-
-    def extract_file_features(self):
-        for feature, va in capa.features.extractors.smda.file.extract_features(self.smda_report, self.path):
-            yield feature, va
-
-    def get_functions(self):
-        for function in self.smda_report.getFunctions():
-            yield function
-
-    def extract_function_features(self, f):
-        for feature, va in capa.features.extractors.smda.function.extract_features(f):
-            yield feature, va
-
-    def get_basic_blocks(self, f):
-        for bb in f.getBlocks():
-            yield bb
-
-    def extract_basic_block_features(self, f, bb):
-        for feature, va in capa.features.extractors.smda.basicblock.extract_features(f, bb):
-            yield feature, va
-
-    def get_instructions(self, f, bb):
-        for smda_ins in bb.getInstructions():
-            yield smda_ins
-
-    def extract_insn_features(self, f, bb, insn):
-        for feature, va in capa.features.extractors.smda.insn.extract_features(f, bb, insn):
-            yield feature, va

capa/features/extractors/smda/basicblock.py

@@ -1,8 +1,7 @@
-import sys
 import string
 import struct
 
-from capa.features import Characteristic
+from capa.features.common import Characteristic
 from capa.features.basicblock import BasicBlock
 from capa.features.extractors.helpers import MIN_STACKSTRING_LEN
 
@@ -15,7 +14,7 @@ def _bb_has_tight_loop(f, bb):
 
 
 def extract_bb_tight_loop(f, bb):
-    """ check basic block for tight loop indicators """
+    """check basic block for tight loop indicators"""
     if _bb_has_tight_loop(f, bb):
         yield Characteristic("tight loop"), bb.offset
 
@@ -39,7 +38,7 @@ def get_operands(smda_ins):
 
 
 def extract_stackstring(f, bb):
-    """ check basic block for stackstring indicators """
+    """check basic block for stackstring indicators"""
     if _bb_has_stackstring(f, bb):
         yield Characteristic("stack string"), bb.offset
 
@@ -117,7 +116,7 @@ def extract_features(f, bb):
       bb (smda.common.SmdaBasicBlock): the basic block to process.
 
     yields:
-      Feature, set[VA]: the features and their location found in this basic block.
+      Tuple[Feature, int]: the features and their location found in this basic block.
     """
     yield BasicBlock(), bb.offset
     for bb_handler in BASIC_BLOCK_HANDLERS:

capa/features/extractors/smda/extractor.py

@@ -0,0 +1,45 @@
+from smda.common.SmdaReport import SmdaReport
+
+import capa.features.extractors.smda.file
+import capa.features.extractors.smda.insn
+import capa.features.extractors.smda.function
+import capa.features.extractors.smda.basicblock
+from capa.features.extractors.base_extractor import FeatureExtractor
+
+
+class SmdaFeatureExtractor(FeatureExtractor):
+    def __init__(self, smda_report: SmdaReport, path):
+        super(SmdaFeatureExtractor, self).__init__()
+        self.smda_report = smda_report
+        self.path = path
+
+    def get_base_address(self):
+        return self.smda_report.base_addr
+
+    def extract_file_features(self):
+        for feature, va in capa.features.extractors.smda.file.extract_features(self.smda_report, self.path):
+            yield feature, va
+
+    def get_functions(self):
+        for function in self.smda_report.getFunctions():
+            yield function
+
+    def extract_function_features(self, f):
+        for feature, va in capa.features.extractors.smda.function.extract_features(f):
+            yield feature, va
+
+    def get_basic_blocks(self, f):
+        for bb in f.getBlocks():
+            yield bb
+
+    def extract_basic_block_features(self, f, bb):
+        for feature, va in capa.features.extractors.smda.basicblock.extract_features(f, bb):
+            yield feature, va
+
+    def get_instructions(self, f, bb):
+        for smda_ins in bb.getInstructions():
+            yield smda_ins
+
+    def extract_insn_features(self, f, bb, insn):
+        for feature, va in capa.features.extractors.smda.insn.extract_features(f, bb, insn):
+            yield feature, va

capa/features/extractors/smda/file.py

@@ -1,65 +1,17 @@
-import struct
-
 # if we have SMDA we definitely have lief
 import lief
 
 import capa.features.extractors.helpers
 import capa.features.extractors.strings
-from capa.features import String, Characteristic
 from capa.features.file import Export, Import, Section
-
-
-def carve(pbytes, offset=0):
-    """
-    Return a list of (offset, size, xor) tuples of embedded PEs
-
-    Based on the version from vivisect:
-    https://github.com/vivisect/vivisect/blob/7be4037b1cecc4551b397f840405a1fc606f9b53/PE/carve.py#L19
-    And its IDA adaptation:
-    capa/features/extractors/ida/file.py
-    """
-    mz_xor = [
-        (
-            capa.features.extractors.helpers.xor_static(b"MZ", i),
-            capa.features.extractors.helpers.xor_static(b"PE", i),
-            i,
-        )
-        for i in range(256)
-    ]
-
-    pblen = len(pbytes)
-    todo = [(pbytes.find(mzx, offset), mzx, pex, i) for mzx, pex, i in mz_xor]
-    todo = [(off, mzx, pex, i) for (off, mzx, pex, i) in todo if off != -1]
-
-    while len(todo):
-
-        off, mzx, pex, i = todo.pop()
-
-        # The MZ header has one field we will check
-        # e_lfanew is at 0x3c
-        e_lfanew = off + 0x3C
-        if pblen < (e_lfanew + 4):
-            continue
-
-        newoff = struct.unpack("<I", capa.features.extractors.helpers.xor_static(pbytes[e_lfanew : e_lfanew + 4], i))[0]
-
-        nextres = pbytes.find(mzx, off + 1)
-        if nextres != -1:
-            todo.append((nextres, mzx, pex, i))
-
-        peoff = off + newoff
-        if pblen < (peoff + 2):
-            continue
-
-        if pbytes[peoff : peoff + 2] == pex:
-            yield (off, i)
+from capa.features.common import String, Characteristic
 
 
 def extract_file_embedded_pe(smda_report, file_path):
     with open(file_path, "rb") as f:
         fbytes = f.read()
 
-    for offset, i in carve(fbytes, 1):
+    for offset, i in capa.features.extractors.helpers.carve_pe(fbytes, 1):
         yield Characteristic("embedded pe"), offset
 
 
@@ -79,8 +31,8 @@ def extract_file_import_names(smda_report, file_path):
         library_name = imported_library.name.lower()
         library_name = library_name[:-4] if library_name.endswith(".dll") else library_name
         for func in imported_library.entries:
+            va = func.iat_address + smda_report.base_addr
             if func.name:
-                va = func.iat_address + smda_report.base_addr
                 for name in capa.features.extractors.helpers.generate_symbols(library_name, func.name):
                     yield Import(name), va
             elif func.is_ordinal:
@@ -112,6 +64,16 @@ def extract_file_strings(smda_report, file_path):
         yield String(s.s), s.offset
 
 
+def extract_file_function_names(smda_report, file_path):
+    """
+    extract the names of statically-linked library functions.
+    """
+    if False:
+        # using a `yield` here to force this to be a generator, not function.
+        yield NotImplementedError("SMDA doesn't have library matching")
+    return
+
+
 def extract_features(smda_report, file_path):
     """
     extract file features from given workspace
@@ -125,7 +87,6 @@ def extract_features(smda_report, file_path):
     """
 
     for file_handler in FILE_HANDLERS:
-        result = file_handler(smda_report, file_path)
         for feature, va in file_handler(smda_report, file_path):
             yield feature, va
 
@@ -136,4 +97,5 @@ FILE_HANDLERS = (
     extract_file_import_names,
     extract_file_section_names,
     extract_file_strings,
+    extract_file_function_names,
 )

capa/features/extractors/smda/function.py

@@ -1,4 +1,4 @@
-from capa.features import Characteristic
+from capa.features.common import Characteristic
 from capa.features.extractors import loops
 
 
@@ -28,7 +28,7 @@ def extract_features(f):
       f (smda.common.SmdaFunction): the function from which to extract features
 
     yields:
-      Feature, set[VA]: the features and their location found in this function.
+      Tuple[Feature, int]: the features and their location found in this function.
     """
     for func_handler in FUNCTION_HANDLERS:
         for feature, va in func_handler(f):

capa/features/extractors/smda/insn.py

@@ -5,7 +5,8 @@ import struct
 from smda.common.SmdaReport import SmdaReport
 
 import capa.features.extractors.helpers
-from capa.features import (
+from capa.features.insn import API, Number, Offset, Mnemonic
+from capa.features.common import (
     ARCH_X32,
     ARCH_X64,
     MAX_BYTES_FEATURE_SIZE,
@@ -14,7 +15,6 @@ from capa.features import (
     String,
     Characteristic,
 )
-from capa.features.insn import API, Number, Offset, Mnemonic
 
 # security cookie checks may perform non-zeroing XORs, these are expected within a certain
 # byte range within the first and returning basic blocks, this helps to reduce FP features
@@ -97,7 +97,7 @@ def read_bytes(smda_report, va, num_bytes=None):
 
     rva = va - smda_report.base_addr
     if smda_report.buffer is None:
-        return
+        raise ValueError("buffer is empty")
     buffer_end = len(smda_report.buffer)
     max_bytes = num_bytes if num_bytes is not None else MAX_BYTES_FEATURE_SIZE
     if rva + max_bytes > buffer_end:
@@ -293,7 +293,7 @@ def extract_insn_peb_access_characteristic_features(f, bb, insn):
 
 
 def extract_insn_segment_access_features(f, bb, insn):
-    """ parse the instruction for access to fs or gs """
+    """parse the instruction for access to fs or gs"""
     operands = [o.strip() for o in insn.operands.split(",")]
     for operand in operands:
         if "fs:" in operand:
@@ -370,7 +370,7 @@ def extract_features(f, bb, insn):
       insn (smda.common.SmdaInstruction): the instruction to process.
 
     yields:
-      Feature, set[VA]: the features and their location found in this insn.
+      Tuple[Feature, int]: the features and their location found in this insn.
     """
     for insn_handler in INSTRUCTION_HANDLERS:
         for feature, va in insn_handler(f, bb, insn):

capa/features/extractors/viv/__init__.py

@@ -1,81 +0,0 @@
-# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
-# Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at: [package root]/LICENSE.txt
-# Unless required by applicable law or agreed to in writing, software distributed under the License
-#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and limitations under the License.
-
-import types
-
-import viv_utils
-
-import capa.features.extractors
-import capa.features.extractors.viv.file
-import capa.features.extractors.viv.insn
-import capa.features.extractors.viv.function
-import capa.features.extractors.viv.basicblock
-from capa.features.extractors import FeatureExtractor
-
-__all__ = ["file", "function", "basicblock", "insn"]
-
-
-def get_va(self):
-    try:
-        # vivisect type
-        return self.va
-    except AttributeError:
-        pass
-
-    raise TypeError()
-
-
-def add_va_int_cast(o):
-    """
-    dynamically add a cast-to-int (`__int__`) method to the given object
-    that returns the value of the `.va` property.
-
-    this bit of skullduggery lets use cast viv-utils objects as ints.
-    the correct way of doing this is to update viv-utils (or subclass the objects here).
-    """
-    setattr(o, "__int__", types.MethodType(get_va, o))
-    return o
-
-
-class VivisectFeatureExtractor(FeatureExtractor):
-    def __init__(self, vw, path):
-        super(VivisectFeatureExtractor, self).__init__()
-        self.vw = vw
-        self.path = path
-
-    def get_base_address(self):
-        # assume there is only one file loaded into the vw
-        return list(self.vw.filemeta.values())[0]["imagebase"]
-
-    def extract_file_features(self):
-        for feature, va in capa.features.extractors.viv.file.extract_features(self.vw, self.path):
-            yield feature, va
-
-    def get_functions(self):
-        for va in sorted(self.vw.getFunctions()):
-            yield add_va_int_cast(viv_utils.Function(self.vw, va))
-
-    def extract_function_features(self, f):
-        for feature, va in capa.features.extractors.viv.function.extract_features(f):
-            yield feature, va
-
-    def get_basic_blocks(self, f):
-        for bb in f.basic_blocks:
-            yield add_va_int_cast(bb)
-
-    def extract_basic_block_features(self, f, bb):
-        for feature, va in capa.features.extractors.viv.basicblock.extract_features(f, bb):
-            yield feature, va
-
-    def get_instructions(self, f, bb):
-        for insn in bb.instructions:
-            yield add_va_int_cast(insn)
-
-    def extract_insn_features(self, f, bb, insn):
-        for feature, va in capa.features.extractors.viv.insn.extract_features(f, bb, insn):
-            yield feature, va

capa/features/extractors/viv/basicblock.py

@@ -10,9 +10,9 @@ import string
 import struct
 
 import envi
-import vivisect.const
+import envi.archs.i386.disasm
 
-from capa.features import Characteristic
+from capa.features.common import Characteristic
 from capa.features.basicblock import BasicBlock
 from capa.features.extractors.helpers import MIN_STACKSTRING_LEN
 
@@ -37,7 +37,7 @@ def _bb_has_tight_loop(f, bb):
     """
     if len(bb.instructions) > 0:
         for bva, bflags in bb.instructions[-1].getBranches():
-            if bflags & vivisect.envi.BR_COND:
+            if bflags & envi.BR_COND:
                 if bva == bb.va:
                     return True
 
@@ -45,7 +45,7 @@ def _bb_has_tight_loop(f, bb):
 
 
 def extract_bb_tight_loop(f, bb):
-    """ check basic block for tight loop indicators """
+    """check basic block for tight loop indicators"""
     if _bb_has_tight_loop(f, bb):
         yield Characteristic("tight loop"), bb.va
 
@@ -68,12 +68,12 @@ def _bb_has_stackstring(f, bb):
 
 
 def extract_stackstring(f, bb):
-    """ check basic block for stackstring indicators """
+    """check basic block for stackstring indicators"""
     if _bb_has_stackstring(f, bb):
         yield Characteristic("stack string"), bb.va
 
 
-def is_mov_imm_to_stack(instr):
+def is_mov_imm_to_stack(instr: envi.archs.i386.disasm.i386Opcode) -> bool:
     """
     Return if instruction moves immediate onto stack
     """
@@ -105,7 +105,7 @@ def is_mov_imm_to_stack(instr):
     return True
 
 
-def get_printable_len(oper):
+def get_printable_len(oper: envi.archs.i386.disasm.i386ImmOper) -> int:
     """
     Return string length if all operand bytes are ascii or utf16-le printable
     """
@@ -117,14 +117,18 @@ def get_printable_len(oper):
         chars = struct.pack("<I", oper.imm)
     elif oper.tsize == 8:
         chars = struct.pack("<Q", oper.imm)
+    else:
+        raise ValueError("unexpected oper.tsize: %d" % (oper.tsize))
+
     if is_printable_ascii(chars):
         return oper.tsize
-    if is_printable_utf16le(chars):
+    elif is_printable_utf16le(chars):
         return oper.tsize / 2
-    return 0
+    else:
+        return 0
 
 
-def is_printable_ascii(chars):
+def is_printable_ascii(chars: bytes) -> bool:
     try:
         chars_str = chars.decode("ascii")
     except UnicodeDecodeError:
@@ -133,9 +137,10 @@ def is_printable_ascii(chars):
         return all(c in string.printable for c in chars_str)
 
 
-def is_printable_utf16le(chars):
+def is_printable_utf16le(chars: bytes) -> bool:
     if all(c == b"\x00" for c in chars[1::2]):
         return is_printable_ascii(chars[::2])
+    return False
 
 
 def extract_features(f, bb):
@@ -147,7 +152,7 @@ def extract_features(f, bb):
       bb (viv_utils.BasicBlock): the basic block to process.
 
     yields:
-      Feature, set[VA]: the features and their location found in this basic block.
+      Tuple[Feature, int]: the features and their location found in this basic block.
     """
     yield BasicBlock(), bb.va
     for bb_handler in BASIC_BLOCK_HANDLERS:

capa/features/extractors/viv/extractor.py

@@ -0,0 +1,76 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+import logging
+
+import viv_utils
+import viv_utils.flirt
+
+import capa.features.extractors.viv.file
+import capa.features.extractors.viv.insn
+import capa.features.extractors.viv.function
+import capa.features.extractors.viv.basicblock
+from capa.features.extractors.base_extractor import FeatureExtractor
+
+logger = logging.getLogger(__name__)
+
+
+class InstructionHandle:
+    """this acts like a vivisect.Opcode but with an __int__() method"""
+
+    def __init__(self, inner):
+        self._inner = inner
+
+    def __int__(self):
+        return self.va
+
+    def __getattr__(self, name):
+        return getattr(self._inner, name)
+
+
+class VivisectFeatureExtractor(FeatureExtractor):
+    def __init__(self, vw, path):
+        super(VivisectFeatureExtractor, self).__init__()
+        self.vw = vw
+        self.path = path
+
+    def get_base_address(self):
+        # assume there is only one file loaded into the vw
+        return list(self.vw.filemeta.values())[0]["imagebase"]
+
+    def extract_file_features(self):
+        for feature, va in capa.features.extractors.viv.file.extract_features(self.vw, self.path):
+            yield feature, va
+
+    def get_functions(self):
+        for va in sorted(self.vw.getFunctions()):
+            yield viv_utils.Function(self.vw, va)
+
+    def extract_function_features(self, f):
+        for feature, va in capa.features.extractors.viv.function.extract_features(f):
+            yield feature, va
+
+    def get_basic_blocks(self, f):
+        return f.basic_blocks
+
+    def extract_basic_block_features(self, f, bb):
+        for feature, va in capa.features.extractors.viv.basicblock.extract_features(f, bb):
+            yield feature, va
+
+    def get_instructions(self, f, bb):
+        for insn in bb.instructions:
+            yield InstructionHandle(insn)
+
+    def extract_insn_features(self, f, bb, insn):
+        for feature, va in capa.features.extractors.viv.insn.extract_features(f, bb, insn):
+            yield feature, va
+
+    def is_library_function(self, va):
+        return viv_utils.flirt.is_library_function(self.vw, va)
+
+    def get_function_name(self, va):
+        return viv_utils.get_function_name(self.vw, va)

capa/features/extractors/viv/file.py

@@ -7,11 +7,14 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 import PE.carve as pe_carve  # vivisect PE
+import viv_utils
+import viv_utils.flirt
 
+import capa.features.insn
 import capa.features.extractors.helpers
 import capa.features.extractors.strings
-from capa.features import String, Characteristic
-from capa.features.file import Export, Import, Section
+from capa.features.file import Export, Import, Section, FunctionName
+from capa.features.common import String, Characteristic
 
 
 def extract_file_embedded_pe(vw, file_path):
@@ -47,7 +50,7 @@ def extract_file_import_names(vw, file_path):
             yield Import(name), va
 
 
-def is_viv_ord_impname(impname):
+def is_viv_ord_impname(impname: str) -> bool:
     """
     return if import name matches vivisect's ordinal naming scheme `'ord%d' % ord`
     """
@@ -80,6 +83,16 @@ def extract_file_strings(vw, file_path):
         yield String(s.s), s.offset
 
 
+def extract_file_function_names(vw, file_path):
+    """
+    extract the names of statically-linked library functions.
+    """
+    for va in sorted(vw.getFunctions()):
+        if viv_utils.flirt.is_library_function(vw, va):
+            name = viv_utils.get_function_name(vw, va)
+            yield FunctionName(name), va
+
+
 def extract_features(vw, file_path):
     """
     extract file features from given workspace
@@ -103,4 +116,5 @@ FILE_HANDLERS = (
     extract_file_import_names,
     extract_file_section_names,
     extract_file_strings,
+    extract_file_function_names,
 )

capa/features/extractors/viv/function.py

@@ -6,9 +6,10 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
+import envi
 import vivisect.const
 
-from capa.features import Characteristic
+from capa.features.common import Characteristic
 from capa.features.extractors import loops
 
 
@@ -41,9 +42,9 @@ def extract_function_loop(f):
             for bva, bflags in bb.instructions[-1].getBranches():
                 # vivisect does not set branch flags for non-conditional jmp so add explicit check
                 if (
-                    bflags & vivisect.envi.BR_COND
-                    or bflags & vivisect.envi.BR_FALL
-                    or bflags & vivisect.envi.BR_TABLE
+                    bflags & envi.BR_COND
+                    or bflags & envi.BR_FALL
+                    or bflags & envi.BR_TABLE
                     or bb.instructions[-1].mnem == "jmp"
                 ):
                     edges.append((bb.va, bva))
@@ -60,7 +61,7 @@ def extract_features(f):
       f (viv_utils.Function): the function from which to extract features
 
     yields:
-      Feature, set[VA]: the features and their location found in this function.
+      Tuple[Feature, int]: the features and their location found in this function.
     """
     for func_handler in FUNCTION_HANDLERS:
         for feature, va in func_handler(f):

capa/features/extractors/viv/helpers.py

@@ -5,10 +5,13 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
+from typing import Optional
+
+from vivisect import VivWorkspace
 from vivisect.const import XR_TO, REF_CODE
 
 
-def get_coderef_from(vw, va):
+def get_coderef_from(vw: VivWorkspace, va: int) -> Optional[int]:
     """
     return first code `tova` whose origin is the specified va
     return None if no code reference is found

capa/features/extractors/viv/indirect_calls.py

@@ -7,11 +7,16 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 import collections
+from typing import TYPE_CHECKING, Set, List, Deque, Tuple, Union, Optional
 
 import envi
 import vivisect.const
 import envi.archs.i386.disasm
 import envi.archs.amd64.disasm
+from vivisect import VivWorkspace
+
+if TYPE_CHECKING:
+    from capa.features.extractors.viv.extractor import InstructionHandle
 
 # pull out consts for lookup performance
 i386RegOper = envi.archs.i386.disasm.i386RegOper
@@ -26,7 +31,7 @@ FAR_BRANCH_MASK = envi.BR_PROC | envi.BR_DEREF | envi.BR_ARCH
 DESTRUCTIVE_MNEMONICS = ("mov", "lea", "pop", "xor")
 
 
-def get_previous_instructions(vw, va):
+def get_previous_instructions(vw: VivWorkspace, va: int) -> List[int]:
     """
     collect the instructions that flow to the given address, local to the current function.
 
@@ -43,12 +48,14 @@ def get_previous_instructions(vw, va):
     # ensure that it fallsthrough to this one.
     loc = vw.getPrevLocation(va, adjacent=True)
     if loc is not None:
-        # from vivisect.const:
-        # location: (L_VA, L_SIZE, L_LTYPE, L_TINFO)
-        (pva, _, ptype, pinfo) = vw.getPrevLocation(va, adjacent=True)
+        ploc = vw.getPrevLocation(va, adjacent=True)
+        if ploc is not None:
+            # from vivisect.const:
+            # location: (L_VA, L_SIZE, L_LTYPE, L_TINFO)
+            (pva, _, ptype, pinfo) = ploc
 
-        if ptype == LOC_OP and not (pinfo & IF_NOFALL):
-            ret.append(pva)
+            if ptype == LOC_OP and not (pinfo & IF_NOFALL):
+                ret.append(pva)
 
     # find any code refs, e.g. jmp, to this location.
     # ignore any calls.
@@ -67,7 +74,7 @@ class NotFoundError(Exception):
     pass
 
 
-def find_definition(vw, va, reg):
+def find_definition(vw: VivWorkspace, va: int, reg: int) -> Tuple[int, Union[int, None]]:
     """
     scan backwards from the given address looking for assignments to the given register.
     if a constant, return that value.
@@ -83,8 +90,8 @@ def find_definition(vw, va, reg):
     raises:
       NotFoundError: when the definition cannot be found.
     """
-    q = collections.deque()
-    seen = set([])
+    q = collections.deque()  # type: Deque[int]
+    seen = set([])  # type: Set[int]
 
     q.extend(get_previous_instructions(vw, va))
     while q:
@@ -128,14 +135,16 @@ def find_definition(vw, va, reg):
     raise NotFoundError()
 
 
-def is_indirect_call(vw, va, insn=None):
+def is_indirect_call(vw: VivWorkspace, va: int, insn: Optional["InstructionHandle"] = None) -> bool:
     if insn is None:
         insn = vw.parseOpcode(va)
 
     return insn.mnem in ("call", "jmp") and isinstance(insn.opers[0], envi.archs.i386.disasm.i386RegOper)
 
 
-def resolve_indirect_call(vw, va, insn=None):
+def resolve_indirect_call(
+    vw: VivWorkspace, va: int, insn: Optional["InstructionHandle"] = None
+) -> Tuple[int, Optional[int]]:
     """
     inspect the given indirect call instruction and attempt to resolve the target address.
 

capa/features/extractors/viv/insn.py

@@ -5,13 +5,20 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-
+import envi
+import envi.exc
+import viv_utils
 import envi.memory
+import viv_utils.flirt
+import envi.archs.i386.regs
+import envi.archs.amd64.regs
 import envi.archs.i386.disasm
+import envi.archs.amd64.disasm
 
 import capa.features.extractors.helpers
 import capa.features.extractors.viv.helpers
-from capa.features import (
+from capa.features.insn import API, Number, Offset, Mnemonic
+from capa.features.common import (
     ARCH_X32,
     ARCH_X64,
     MAX_BYTES_FEATURE_SIZE,
@@ -20,7 +27,6 @@ from capa.features import (
     String,
     Characteristic,
 )
-from capa.features.insn import API, Number, Offset, Mnemonic
 from capa.features.extractors.viv.indirect_calls import NotFoundError, resolve_indirect_call
 
 # security cookie checks may perform non-zeroing XORs, these are expected within a certain
@@ -74,7 +80,6 @@ def extract_insn_api_features(f, bb, insn):
     # example:
     #
     #    call dword [0x00473038]
-
     if insn.mnem not in ("call", "jmp"):
         return
 
@@ -96,7 +101,7 @@ def extract_insn_api_features(f, bb, insn):
     # call via thunk on x86,
     # see 9324d1a8ae37a36ae560c37448c9705a at 0x407985
     #
-    # this is also how calls to internal functions may be decoded on x64.
+    # this is also how calls to internal functions may be decoded on x32 and x64.
     # see Lab21-01.exe_:0x140001178
     #
     # follow chained thunks, e.g. in 82bf6347acf15e5d883715dc289d8a2b at 0x14005E0FF in
@@ -111,6 +116,11 @@ def extract_insn_api_features(f, bb, insn):
         if not target:
             return
 
+        if viv_utils.flirt.is_library_function(f.vw, target):
+            name = viv_utils.get_function_name(f.vw, target)
+            yield API(name), insn.va
+            return
+
         for _ in range(THUNK_CHAIN_DEPTH_DELTA):
             if target in imports:
                 dll, symbol = imports[target]
@@ -171,7 +181,7 @@ def extract_insn_number_features(f, bb, insn):
             # assume its not also a constant.
             continue
 
-        if insn.mnem == "add" and insn.opers[0].isReg() and insn.opers[0].reg == envi.archs.i386.disasm.REG_ESP:
+        if insn.mnem == "add" and insn.opers[0].isReg() and insn.opers[0].reg == envi.archs.i386.regs.REG_ESP:
             # skip things like:
             #
             #    .text:00401140                 call    sub_407E2B
@@ -214,7 +224,7 @@ def derefs(vw, p):
         p = next
 
 
-def read_memory(vw, va, size):
+def read_memory(vw, va: int, size: int) -> bytes:
     # as documented in #176, vivisect will not readMemory() when the section is not marked readable.
     #
     # but here, we don't care about permissions.
@@ -227,10 +237,10 @@ def read_memory(vw, va, size):
             mva, msize, mperms, mfname = mmap
             offset = va - mva
             return mbytes[offset : offset + size]
-    raise envi.SegmentationViolation(va)
+    raise envi.exc.SegmentationViolation(va)
 
 
-def read_bytes(vw, va):
+def read_bytes(vw, va: int) -> bytes:
     """
     read up to MAX_BYTES_FEATURE_SIZE from the given address.
 
@@ -239,7 +249,7 @@ def read_bytes(vw, va):
     """
     segm = vw.getSegment(va)
     if not segm:
-        raise envi.SegmentationViolation(va)
+        raise envi.exc.SegmentationViolation(va)
 
     segm_end = segm[0] + segm[1]
     try:
@@ -248,7 +258,7 @@ def read_bytes(vw, va):
             return read_memory(vw, va, segm_end - va)
         else:
             return read_memory(vw, va, MAX_BYTES_FEATURE_SIZE)
-    except envi.SegmentationViolation:
+    except envi.exc.SegmentationViolation:
         raise
 
 
@@ -280,7 +290,7 @@ def extract_insn_bytes_features(f, bb, insn):
         for v in derefs(f.vw, v):
             try:
                 buf = read_bytes(f.vw, v)
-            except envi.SegmentationViolation:
+            except envi.exc.SegmentationViolation:
                 continue
 
             if capa.features.extractors.helpers.all_zeros(buf):
@@ -289,10 +299,10 @@ def extract_insn_bytes_features(f, bb, insn):
             yield Bytes(buf), insn.va
 
 
-def read_string(vw, offset):
+def read_string(vw, offset: int) -> str:
     try:
         alen = vw.detectString(offset)
-    except envi.SegmentationViolation:
+    except envi.exc.SegmentationViolation:
         pass
     else:
         if alen > 0:
@@ -300,7 +310,7 @@ def read_string(vw, offset):
 
     try:
         ulen = vw.detectUnicode(offset)
-    except envi.SegmentationViolation:
+    except envi.exc.SegmentationViolation:
         pass
     except IndexError:
         # potential vivisect bug detecting Unicode at segment end
@@ -361,14 +371,14 @@ def extract_insn_offset_features(f, bb, insn):
         #       reg   ^
         #             disp
         if isinstance(oper, envi.archs.i386.disasm.i386RegMemOper):
-            if oper.reg == envi.archs.i386.disasm.REG_ESP:
+            if oper.reg == envi.archs.i386.regs.REG_ESP:
                 continue
 
-            if oper.reg == envi.archs.i386.disasm.REG_EBP:
+            if oper.reg == envi.archs.i386.regs.REG_EBP:
                 continue
 
             # TODO: do x64 support for real.
-            if oper.reg == envi.archs.amd64.disasm.REG_RBP:
+            if oper.reg == envi.archs.amd64.regs.REG_RBP:
                 continue
 
             # viv already decodes offsets as signed
@@ -389,18 +399,18 @@ def extract_insn_offset_features(f, bb, insn):
             yield Offset(v, arch=get_arch(f.vw)), insn.va
 
 
-def is_security_cookie(f, bb, insn):
+def is_security_cookie(f, bb, insn) -> bool:
     """
     check if an instruction is related to security cookie checks
     """
     # security cookie check should use SP or BP
     oper = insn.opers[1]
     if oper.isReg() and oper.reg not in [
-        envi.archs.i386.disasm.REG_ESP,
-        envi.archs.i386.disasm.REG_EBP,
+        envi.archs.i386.regs.REG_ESP,
+        envi.archs.i386.regs.REG_EBP,
         # TODO: do x64 support for real.
-        envi.archs.amd64.disasm.REG_RBP,
-        envi.archs.amd64.disasm.REG_RSP,
+        envi.archs.amd64.regs.REG_RBP,
+        envi.archs.amd64.regs.REG_RSP,
     ]:
         return False
 
@@ -476,7 +486,7 @@ def extract_insn_peb_access_characteristic_features(f, bb, insn):
 
 
 def extract_insn_segment_access_features(f, bb, insn):
-    """ parse the instruction for access to fs or gs """
+    """parse the instruction for access to fs or gs"""
     prefix = insn.getPrefixName()
 
     if prefix == "fs":
@@ -486,7 +496,7 @@ def extract_insn_segment_access_features(f, bb, insn):
         yield Characteristic("gs access"), insn.va
 
 
-def get_section(vw, va):
+def get_section(vw, va: int):
     for start, length, _, __ in vw.getMemoryMaps():
         if start <= va < start + length:
             return start
@@ -597,7 +607,7 @@ def extract_features(f, bb, insn):
       insn (vivisect...Instruction): the instruction to process.
 
     yields:
-      Feature, set[VA]: the features and their location found in this insn.
+      Tuple[Feature, int]: the features and their location found in this insn.
     """
     for insn_handler in INSTRUCTION_HANDLERS:
         for feature, va in insn_handler(f, bb, insn):

capa/features/file.py

@@ -6,22 +6,33 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-from capa.features import Feature
+from capa.features.common import Feature
 
 
 class Export(Feature):
-    def __init__(self, value, description=None):
+    def __init__(self, value: str, description=None):
         # value is export name
         super(Export, self).__init__(value, description=description)
 
 
 class Import(Feature):
-    def __init__(self, value, description=None):
+    def __init__(self, value: str, description=None):
         # value is import name
         super(Import, self).__init__(value, description=description)
 
 
 class Section(Feature):
-    def __init__(self, value, description=None):
+    def __init__(self, value: str, description=None):
         # value is section name
         super(Section, self).__init__(value, description=description)
+
+
+class FunctionName(Feature):
+    """recognized name for statically linked function"""
+
+    def __init__(self, name: str, description=None):
+        # value is function name
+        super(FunctionName, self).__init__(name, description=description)
+        # override the name property set by `capa.features.Feature`
+        # that would be `functionname` (note missing dash)
+        self.name = "function-name"

capa/features/freeze.py

@@ -52,12 +52,13 @@ See the License for the specific language governing permissions and limitations
 import json
 import zlib
 import logging
+import os.path
 
-import capa.features
 import capa.features.file
 import capa.features.insn
+import capa.features.common
 import capa.features.basicblock
-import capa.features.extractors
+import capa.features.extractors.base_extractor
 from capa.helpers import hex
 
 logger = logging.getLogger(__name__)
@@ -67,7 +68,7 @@ def serialize_feature(feature):
     return feature.freeze_serialize()
 
 
-KNOWN_FEATURES = {F.__name__: F for F in capa.features.Feature.__subclasses__()}
+KNOWN_FEATURES = {F.__name__: F for F in capa.features.common.Feature.__subclasses__()}
 
 
 def deserialize_feature(doc):
@@ -80,7 +81,7 @@ def dumps(extractor):
     serialize the given extractor to a string
 
     args:
-      extractor: capa.features.extractor.FeatureExtractor:
+      extractor: capa.features.extractors.base_extractor.FeatureExtractor:
 
     returns:
       str: the serialized features.
@@ -122,7 +123,7 @@ def dumps(extractor):
                 )
 
             for insnva, insn in sorted(
-                [(insn.__int__(), insn) for insn in extractor.get_instructions(f, bb)], key=lambda p: p[0]
+                [(int(insn), insn) for insn in extractor.get_instructions(f, bb)], key=lambda p: p[0]
             ):
                 ret["functions"][hex(f)][hex(bb)].append(hex(insnva))
 
@@ -217,7 +218,7 @@ def loads(s):
         feature = deserialize_feature(feature[:2])
         features["functions"][loc[0]]["basic blocks"][loc[1]]["instructions"][loc[2]]["features"].append((va, feature))
 
-    return capa.features.extractors.NullFeatureExtractor(features)
+    return capa.features.extractors.base_extractor.NullFeatureExtractor(features)
 
 
 MAGIC = "capa0000".encode("ascii")
@@ -228,7 +229,7 @@ def dump(extractor):
     return MAGIC + zlib.compress(dumps(extractor).encode("utf-8"))
 
 
-def is_freeze(buf):
+def is_freeze(buf: bytes) -> bool:
     return buf[: len(MAGIC)] == MAGIC
 
 
@@ -248,45 +249,16 @@ def main(argv=None):
     if argv is None:
         argv = sys.argv[1:]
 
-    formats = [
-        ("auto", "(default) detect file type automatically"),
-        ("pe", "Windows PE file"),
-        ("sc32", "32-bit shellcode"),
-        ("sc64", "64-bit shellcode"),
-    ]
-    format_help = ", ".join(["%s: %s" % (f[0], f[1]) for f in formats])
-
     parser = argparse.ArgumentParser(description="save capa features to a file")
-    parser.add_argument("sample", type=str, help="Path to sample to analyze")
+    capa.main.install_common_args(parser, {"sample", "format", "backend", "signatures"})
     parser.add_argument("output", type=str, help="Path to output file")
-    parser.add_argument("-v", "--verbose", action="store_true", help="Enable verbose output")
-    parser.add_argument("-q", "--quiet", action="store_true", help="Disable all output but errors")
-    parser.add_argument(
-        "-f", "--format", choices=[f[0] for f in formats], default="auto", help="Select sample format, %s" % format_help
-    )
-    if sys.version_info >= (3, 0):
-        parser.add_argument(
-            "-b",
-            "--backend",
-            type=str,
-            help="select the backend to use",
-            choices=(capa.main.BACKEND_VIV, capa.main.BACKEND_SMDA),
-            default=capa.main.BACKEND_VIV,
-        )
     args = parser.parse_args(args=argv)
+    capa.main.handle_common_args(args)
 
-    if args.quiet:
-        logging.basicConfig(level=logging.ERROR)
-        logging.getLogger().setLevel(logging.ERROR)
-    elif args.verbose:
-        logging.basicConfig(level=logging.DEBUG)
-        logging.getLogger().setLevel(logging.DEBUG)
-    else:
-        logging.basicConfig(level=logging.INFO)
-        logging.getLogger().setLevel(logging.INFO)
+    sigpaths = capa.main.get_signatures(args.signatures)
+
+    extractor = capa.main.get_extractor(args.sample, args.format, args.backend, sigpaths, False)
 
-    backend = args.backend if sys.version_info > (3, 0) else capa.main.BACKEND_VIV
-    extractor = capa.main.get_extractor(args.sample, args.format, backend)
     with open(args.output, "wb") as f:
         f.write(dump(extractor))
 

capa/features/insn.py

@@ -6,11 +6,12 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-from capa.features import Feature
+import capa.render.utils
+from capa.features.common import Feature
 
 
 class API(Feature):
-    def __init__(self, name, description=None):
+    def __init__(self, name: str, description=None):
         # Downcase library name if given
         if "." in name:
             modname, _, impname = name.rpartition(".")
@@ -20,21 +21,21 @@ class API(Feature):
 
 
 class Number(Feature):
-    def __init__(self, value, arch=None, description=None):
+    def __init__(self, value: int, arch=None, description=None):
         super(Number, self).__init__(value, arch=arch, description=description)
 
     def get_value_str(self):
-        return "0x%X" % self.value
+        return capa.render.utils.hex(self.value)
 
 
 class Offset(Feature):
-    def __init__(self, value, arch=None, description=None):
+    def __init__(self, value: int, arch=None, description=None):
         super(Offset, self).__init__(value, arch=arch, description=description)
 
     def get_value_str(self):
-        return "0x%X" % self.value
+        return capa.render.utils.hex(self.value)
 
 
 class Mnemonic(Feature):
-    def __init__(self, value, description=None):
+    def __init__(self, value: str, description=None):
         super(Mnemonic, self).__init__(value, description=description)

capa/helpers.py

@@ -12,23 +12,10 @@ _hex = hex
 
 
 def hex(i):
-    # under py2.7, long integers get formatted with a trailing `L`
-    # and this is not pretty. so strip it out.
-    return _hex(oint(i)).rstrip("L")
+    return _hex(int(i))
 
 
-def oint(i):
-    # there seems to be some trouble with using `int(viv_utils.Function)`
-    # with the black magic we do with binding the `__int__()` routine.
-    # i haven't had a chance to debug this yet (and i have no hotel wifi).
-    # so in the meantime, detect this, and call the method directly.
-    try:
-        return int(i)
-    except TypeError:
-        return i.__int__()
-
-
-def get_file_taste(sample_path):
+def get_file_taste(sample_path: str) -> bytes:
     if not os.path.exists(sample_path):
         raise IOError("sample path %s does not exist or cannot be accessed" % sample_path)
     with open(sample_path, "rb") as f:

capa/ida/helpers.py

@@ -10,11 +10,12 @@ import logging
 import datetime
 
 import idc
-import six
 import idaapi
 import idautils
 
 import capa
+import capa.version
+import capa.features.common
 
 logger = logging.getLogger("capa")
 
@@ -24,6 +25,7 @@ SUPPORTED_IDA_VERSIONS = [
     "7.3",
     "7.4",
     "7.5",
+    "7.6",
 ]
 
 # file type names as returned by idaapi.get_file_type_name()
@@ -71,7 +73,7 @@ def get_disasm_line(va):
 
 
 def is_func_start(ea):
-    """ check if function stat exists at virtual address """
+    """check if function stat exists at virtual address"""
     f = idaapi.get_func(ea)
     return f and f.start_ea == ea
 
@@ -85,16 +87,16 @@ def get_func_start_ea(ea):
 def get_file_md5():
     """ """
     md5 = idautils.GetInputFileMD5()
-    if not isinstance(md5, six.string_types):
-        md5 = capa.features.bytes_to_str(md5)
+    if not isinstance(md5, str):
+        md5 = capa.features.common.bytes_to_str(md5)
     return md5
 
 
 def get_file_sha256():
     """ """
     sha256 = idaapi.retrieve_input_file_sha256()
-    if not isinstance(sha256, six.string_types):
-        sha256 = capa.features.bytes_to_str(sha256)
+    if not isinstance(sha256, str):
+        sha256 = capa.features.common.bytes_to_str(sha256)
     return sha256
 
 

capa/ida/plugin/README.md

@@ -1,13 +1,12 @@
 ![capa explorer](../../../.github/capa-explorer-logo.png)
 
-capa explorer is an IDA Pro plugin written in Python that integrates the FLARE team's open-source framework, capa, with IDA. capa is a framework that uses a well-defined collection of rules to 
+capa explorer is an IDAPython plugin that integrates the FLARE team's open-source framework, capa, with IDA Pro. capa is a framework that uses a well-defined collection of rules to 
 identify capabilities in a program. You can run capa against a PE file or shellcode and it tells you what it thinks the program can do. For example, it might suggest that 
-the program is a backdoor, can install services, or relies on HTTP to communicate. You can use capa explorer to run capa directly on an IDA database without requiring access
-to or execution of the source binary. Once a database has been analyzed, capa explorer can be used to quickly identify and navigate to interesting areas of a program and manually build new capa rules out
-of the features extracted directly from your IDB.
+the program is a backdoor, can install services, or relies on HTTP to communicate. capa explorer runs capa directly against your IDA Pro database (IDB) without requiring access
+to the original binary file. Once a database has been analyzed, capa explorer helps you identify interesting areas of a program and build new capa rules using features extracted from your IDB.
 
 We love using capa explorer during malware analysis because it teaches us what parts of a program suggest a behavior. As we click on rows, capa explorer jumps directly 
-to important addresses in the IDA Pro database and highlights key features in the Disassembly view so they stand out visually. To illustrate, we use capa explorer to 
+to important addresses in the IDB and highlights key features in the Disassembly view so they stand out visually. To illustrate, we use capa explorer to 
 analyze Lab 14-02 from [Practical Malware Analysis](https://nostarch.com/malware) (PMA) available [here](https://practicalmalwareanalysis.com/labs/). Our goal is to understand 
 the program's functionality.
 
@@ -15,16 +14,15 @@ After loading Lab 14-02 into IDA and analyzing the database with capa explorer,
 
 ![](../../../doc/img/explorer_condensed.png)
 
-We can use capa explorer to navigate the IDA Disassembly view directly to the suspect function and get an assembly-level breakdown of why capa matched `self delete via COMSPEC environment variable` 
-for this particular function.
+We can use capa explorer to navigate our Disassembly view directly to the suspect function and get an assembly-level breakdown of why capa matched `self delete via COMSPEC environment variable`.
 
 ![](../../../doc/img/explorer_expanded.png)
 
 Using the `Rule Information` and `Details` columns capa explorer shows us that the suspect function matched `self delete via COMSPEC environment variable` because it contains capa rule matches for `create process`, `get COMSPEC environment variable`,
-and `query environment variable`, references to the strings `COMSPEC`, ` > nul`, and `/c del`, and calls to the Windows API functions `GetEnvironmentVariableA` and `ShellExecuteEx`.
+and `query environment variable`, references to the strings `COMSPEC`, ` > nul`, and `/c del `, and calls to the Windows API functions `GetEnvironmentVariableA` and `ShellExecuteEx`.
 
-capa explorer also helps you build new capa rules. To start select the `Rule Generator` tab, navigate to a function in the IDA `Disassembly` view,
-and click `Analyze`. capa explorer will extract features from this function and display them in the `Function Features` pane. You can add features listed in this pane to the `Editor` pane
+capa explorer also helps you build new capa rules. To start select the `Rule Generator` tab, navigate to a function in your Disassembly view,
+and click `Analyze`. capa explorer will extract features from the function and display them in the `Features` pane. You can add features listed in this pane to the `Editor` pane
 by either double-clicking a feature or using multi-select + right-click to add multiple features at once. The `Preview` and `Editor` panes help edit your rule. Use the `Preview` pane
 to modify the rule text directly and the `Editor` pane to construct and rearrange your hierarchy of statements and features. When you finish a rule you can save it directly to a file by clicking `Save`.
 
@@ -36,15 +34,32 @@ For more information on the FLARE team's open-source framework, capa, check out
 
 ### Requirements
 
-capa explorer supports the following IDA setups:
+capa explorer supports Python >= 3.6 and the following IDA Pro versions:
 
-* IDA Pro 7.4+ with Python 2.7 or Python 3.
+* IDA 7.4
+* IDA 7.5
+* IDA 7.6 (caveat below)
 
 If you encounter issues with your specific setup, please open a new [Issue](https://github.com/fireeye/capa/issues).
 
+#### IDA 7.6 caveat: IDA 7.6sp1 or patch required
+
+As described [here](https://www.hex-rays.com/blog/ida-7-6-empty-qtreeview-qtreewidget/):
+
+> A rather nasty issue evaded our testing and found its way into IDA 7.6: using the PyQt5 modules that are shipped with IDA, QTreeView (or QTreeWidget) instances will always fail to display contents.
+
+Therefore, in order to use capa under IDA 7.6 you need the [Service Pack 1 for IDA 7.6](https://www.hex-rays.com/products/ida/news/7_6sp1). Alternatively, you can download and install the fix corresponding to your IDA installation, replacing the original QtWidgets DLL with the one contained in the .zip file (links to Hex-Rays):
+
+
+  - Windows: [pyqt5_qtwidgets_win](https://www.hex-rays.com/wp-content/uploads/2021/04/pyqt5_qtwidgets_win.zip)
+  - Linux: [pyqt5_qtwidgets_linux](https://www.hex-rays.com/wp-content/uploads/2021/04/pyqt5_qtwidgets_linux.zip)
+  - MacOS (Intel): [pyqt5_qtwidgets_mac_x64](https://www.hex-rays.com/wp-content/uploads/2021/04/pyqt5_qtwidgets_mac_x64.zip)
+  - MacOS (AppleSilicon): [pyqt5_qtwidgets_mac_arm](https://www.hex-rays.com/wp-content/uploads/2021/04/pyqt5_qtwidgets_mac_arm.zip)
+
+
 ### Supported File Types
 
-capa explorer is limited to the file types supported by capa, which includes:
+capa explorer is limited to the file types supported by capa, which include:
 
 * Windows 32-bit and 64-bit PE files
 * Windows 32-bit and 64-bit shellcode
@@ -62,50 +77,49 @@ You can install capa explorer using the following steps:
 
 ### Usage
 
-1. Run IDA and analyze a supported file type (select the `Manual Load` and `Load Resources` options in IDA for best results)
+1. Open IDA and analyze a supported file type (select the `Manual Load` and `Load Resources` options in IDA for best results)
 2. Open capa explorer in IDA by navigating to `Edit > Plugins > FLARE capa explorer` or using the keyboard shortcut `Alt+F5`
+   You can also use `ida_loader.load_and_run_plugin("capa_explorer", arg)`. `arg` is a bitflag for which setting the LSB enables automatic analysis. See `capa.ida.plugin.form.Options` for more details.
 3. Select the `Program Analysis` tab
 4. Click the `Analyze` button
 
 When running capa explorer for the first time you are prompted to select a file directory containing capa rules. The plugin conveniently
-remembers your selection for future runs; you can change this selection by navigating to `Settings > Change default rules directory...`. We recommend 
+remembers your selection for future runs; you can change this selection and other default settings by clicking `Settings`. We recommend 
 downloading and using the [standard collection of capa rules](https://github.com/fireeye/capa-rules) when getting started with the plugin.
 
 #### Tips for Program Analysis
 
 * Start analysis by clicking the `Analyze` button
-* Reset the plugin user interface and remove highlighting from IDA disassembly view by clicking the `Reset` button
-* Change your capa rules directory by navigating to `Settings > Change default rules directory...` from the plugin menu
+* Reset the plugin user interface and remove highlighting from your Disassembly view by clicking the `Reset` button
+* Change your capa rules directory and other default settings by clicking `Settings`
 * Hover your cursor over a rule match to view the source content of the rule
-* Double-click the `Address` column to navigate the IDA Disassembly view to the associated feature
+* Double-click the `Address` column to navigate your Disassembly view to the address of the associated feature
 * Double-click a result in the `Rule Information` column to expand its children
-* Select a checkbox in the `Rule Information` column to highlight the address of the associated feature in the IDA Dissasembly view
+* Select a checkbox in the `Rule Information` column to highlight the address of the associated feature in your Dissasembly view
 
 #### Tips for Rule Generator
 
-* Navigate to a function in the `Disassembly` view and click`Analyze` to get started
-* Double-click or multi-select + right-click in the `Function Features` pane to add features to the `Editor` pane
-* Right-click features in the `Editor` pane to make modifications
-* Drag-and-drop (single click + multi-select support) features in the `Editor` pane to quickly build a hierarchy of statements and features
-* Right-click anywhere in the `Editor` pane not on a feature to quickly remove all features
-* Add descriptions/comments by placing editing the appropriate column in the `Editor` pane
-* Directly edit rule text, including rule metadata fields using the `Preview` pane
-* Change the default rule author and default scope displayed in the `Preview` pane by navigating to `Settings`
+* Navigate to a function in your Disassembly view and click`Analyze` to get started
+* Double-click or use multi-select + right-click to add features from the `Features` pane to the `Editor` pane
+* Right-click features in the `Editor` pane to make context-specific modifications
+* Drag-and-drop (single click + multi-select support) features in the `Editor` pane to construct your hierarchy of statements and features
+* Right-click anywhere in the `Editor` pane not on a feature to remove all features
+* Add descriptions or comments to a feature by editing the corresponding column in the `Editor` pane
+* Directly edit rule text and metadata fields using the `Preview` pane
+* Change the default rule author and default rule scope displayed in the `Preview` pane by clicking `Settings`
 
 ## Development
 
-Because capa explorer is packaged with capa you will need to install capa locally for development.
-
-You can install capa locally by following the steps outlined in `Method 3: Inspecting the capa source code` of the [capa 
+capa explorer is packaged with capa so you will need to install capa locally for development. You can install capa locally by following the steps outlined in `Method 3: Inspecting the capa source code` of the [capa 
 installation guide](https://github.com/fireeye/capa/blob/master/doc/installation.md#method-3-inspecting-the-capa-source-code). Once installed, copy [capa_explorer.py](https://raw.githubusercontent.com/fireeye/capa/master/capa/ida/plugin/capa_explorer.py) 
-to your IDA plugins directory to run the plugin in IDA.
+to your plugins directory to install capa explorer in IDA.
 
 ### Components
 
 capa explorer consists of two main components:
 
-* An IDA [feature extractor](https://github.com/fireeye/capa/tree/master/capa/features/extractors/ida) built on top of IDA's binary analysis engine
-  * This component uses IDAPython to extract [capa features](https://github.com/fireeye/capa-rules/blob/master/doc/format.md#extracted-features) from the IDA database such as strings, 
+* An [feature extractor](https://github.com/fireeye/capa/tree/master/capa/features/extractors/ida) built on top of IDA's binary analysis engine
+  * This component uses IDAPython to extract [capa features](https://github.com/fireeye/capa-rules/blob/master/doc/format.md#extracted-features) from your IDBs such as strings, 
 disassembly, and control flow; these extracted features are used by capa to find feature combinations that result in a rule match
 * An [interactive user interface](https://github.com/fireeye/capa/tree/master/capa/ida/plugin) for displaying and exploring capa rule matches
-  * This component integrates the IDA feature extractor and capa, providing an interactive user interface to dissect rule matches found by capa using features extracted by the IDA feature extractor
+  * This component integrates the feature extractor and capa, providing an interactive user interface to dissect rule matches found by capa using features extracted directly from your IDBs

capa/ida/plugin/__init__.py

@@ -11,7 +11,6 @@ import logging
 import idaapi
 import ida_kernwin
 
-from capa.ida.helpers import is_supported_file_type, is_supported_ida_version
 from capa.ida.plugin.form import CapaExplorerForm
 from capa.ida.plugin.icon import ICON
 
@@ -41,10 +40,12 @@ class CapaExplorerPlugin(idaapi.plugin_t):
         """called when IDA is loading the plugin"""
         logging.basicConfig(level=logging.INFO)
 
+        import capa.ida.helpers
+
         # do not load plugin if IDA version/file type not supported
-        if not is_supported_ida_version():
+        if not capa.ida.helpers.is_supported_ida_version():
             return idaapi.PLUGIN_SKIP
-        if not is_supported_file_type():
+        if not capa.ida.helpers.is_supported_file_type():
             return idaapi.PLUGIN_SKIP
         return idaapi.PLUGIN_OK
 
@@ -53,8 +54,14 @@ class CapaExplorerPlugin(idaapi.plugin_t):
         pass
 
     def run(self, arg):
-        """called when IDA is running the plugin as a script"""
-        self.form = CapaExplorerForm(self.PLUGIN_NAME)
+        """
+        called when IDA is running the plugin as a script
+
+        args:
+          arg (int): bitflag. Setting LSB enables automatic analysis upon
+          loading. The other bits are currently undefined. See `form.Options`.
+        """
+        self.form = CapaExplorerForm(self.PLUGIN_NAME, arg)
         return True
 
 

capa/ida/plugin/form.py

@@ -20,9 +20,12 @@ from PyQt5 import QtGui, QtCore, QtWidgets
 
 import capa.main
 import capa.rules
+import capa.engine
 import capa.ida.helpers
-import capa.render.utils as rutils
-import capa.features.extractors.ida
+import capa.render.json
+import capa.features.common
+import capa.render.result_document
+import capa.features.extractors.ida.extractor
 from capa.ida.plugin.icon import QICON
 from capa.ida.plugin.view import (
     CapaExplorerQtreeView,
@@ -37,6 +40,17 @@ from capa.ida.plugin.proxy import CapaExplorerRangeProxyModel, CapaExplorerSearc
 logger = logging.getLogger(__name__)
 settings = ida_settings.IDASettings("capa")
 
+CAPA_SETTINGS_RULE_PATH = "rule_path"
+CAPA_SETTINGS_RULEGEN_AUTHOR = "rulegen_author"
+CAPA_SETTINGS_RULEGEN_SCOPE = "rulegen_scope"
+
+from enum import IntFlag
+
+
+class Options(IntFlag):
+    DEFAULT = 0
+    ANALYZE = 1  # Runs the analysis when starting the explorer
+
 
 def write_file(path, data):
     """ """
@@ -74,7 +88,7 @@ def find_func_features(f, extractor):
                 _bb_features[feature].add(ea)
                 func_features[feature].add(ea)
 
-        bb_features[capa.helpers.oint(bb)] = _bb_features
+        bb_features[int(bb)] = _bb_features
 
     return func_features, bb_features
 
@@ -93,10 +107,10 @@ def find_func_matches(f, ruleset, func_features, bb_features):
         for (name, res) in matches.items():
             bb_matches[name].extend(res)
             for (ea, _) in res:
-                func_features[capa.features.MatchedRule(name)].add(ea)
+                func_features[capa.features.common.MatchedRule(name)].add(ea)
 
     # find rule matches for function, function features include rule matches for basic blocks
-    _, matches = capa.engine.match(ruleset.function_rules, func_features, capa.helpers.oint(f))
+    _, matches = capa.engine.match(ruleset.function_rules, func_features, int(f))
     for (name, res) in matches.items():
         func_matches[name].extend(res)
 
@@ -151,7 +165,7 @@ class CapaExplorerProgressIndicator(QtCore.QObject):
         self.progress.emit("extracting features from %s" % text)
 
 
-class CapaExplorerFeatureExtractor(capa.features.extractors.ida.IdaFeatureExtractor):
+class CapaExplorerFeatureExtractor(capa.features.extractors.ida.extractor.IdaFeatureExtractor):
     """subclass the IdaFeatureExtractor
 
     track progress during feature extraction, also allow user to cancel feature extraction
@@ -166,10 +180,64 @@ class CapaExplorerFeatureExtractor(capa.features.extractors.ida.IdaFeatureExtrac
         return super(CapaExplorerFeatureExtractor, self).extract_function_features(f)
 
 
+class QLineEditClicked(QtWidgets.QLineEdit):
+    def __init__(self, content, parent=None):
+        """ """
+        super(QLineEditClicked, self).__init__(content, parent)
+
+    def mouseReleaseEvent(self, e):
+        """ """
+        old = self.text()
+        new = str(
+            QtWidgets.QFileDialog.getExistingDirectory(
+                self.parent(), "Please select a capa rules directory", settings.user.get(CAPA_SETTINGS_RULE_PATH, "")
+            )
+        )
+        if new:
+            self.setText(new)
+        else:
+            self.setText(old)
+
+
+class CapaSettingsInputDialog(QtWidgets.QDialog):
+    def __init__(self, title, parent=None):
+        """ """
+        super(CapaSettingsInputDialog, self).__init__(parent)
+
+        self.setWindowTitle(title)
+        self.setMinimumWidth(500)
+        self.setWindowFlags(self.windowFlags() & ~QtCore.Qt.WindowContextHelpButtonHint)
+
+        self.edit_rule_path = QLineEditClicked(settings.user.get(CAPA_SETTINGS_RULE_PATH, ""))
+        self.edit_rule_author = QtWidgets.QLineEdit(settings.user.get(CAPA_SETTINGS_RULEGEN_AUTHOR, ""))
+        self.edit_rule_scope = QtWidgets.QComboBox()
+
+        scopes = ("file", "function", "basic block")
+
+        self.edit_rule_scope.addItems(scopes)
+        self.edit_rule_scope.setCurrentIndex(scopes.index(settings.user.get(CAPA_SETTINGS_RULEGEN_SCOPE, "function")))
+
+        buttons = QtWidgets.QDialogButtonBox(QtWidgets.QDialogButtonBox.Ok | QtWidgets.QDialogButtonBox.Cancel, self)
+
+        layout = QtWidgets.QFormLayout(self)
+        layout.addRow("capa rules path", self.edit_rule_path)
+        layout.addRow("Default rule author", self.edit_rule_author)
+        layout.addRow("Default rule scope", self.edit_rule_scope)
+
+        layout.addWidget(buttons)
+
+        buttons.accepted.connect(self.accept)
+        buttons.rejected.connect(self.reject)
+
+    def get_values(self):
+        """ """
+        return self.edit_rule_path.text(), self.edit_rule_author.text(), self.edit_rule_scope.currentText()
+
+
 class CapaExplorerForm(idaapi.PluginForm):
     """form element for plugin interface"""
 
-    def __init__(self, name):
+    def __init__(self, name, option=Options.DEFAULT):
         """initialize form elements"""
         super(CapaExplorerForm, self).__init__()
 
@@ -197,11 +265,11 @@ class CapaExplorerForm(idaapi.PluginForm):
         self.view_rulegen = None
         self.view_tabs = None
         self.view_tab_rulegen = None
-        self.view_menu_bar = None
         self.view_status_label = None
         self.view_buttons = None
         self.view_analyze_button = None
         self.view_reset_button = None
+        self.view_settings_button = None
         self.view_save_button = None
 
         self.view_rulegen_preview = None
@@ -217,6 +285,9 @@ class CapaExplorerForm(idaapi.PluginForm):
 
         self.Show()
 
+        if (option & Options.ANALYZE) == Options.ANALYZE:
+            self.analyze_program()
+
     def OnCreate(self, form):
         """called when plugin form is created
 
@@ -273,10 +344,6 @@ class CapaExplorerForm(idaapi.PluginForm):
         self.load_view_status_label()
         self.load_view_buttons()
 
-        # load menu bar and sub menus
-        self.load_view_menu_bar()
-        self.load_configure_menu()
-
         # load parent view
         self.load_view_parent()
 
@@ -285,11 +352,6 @@ class CapaExplorerForm(idaapi.PluginForm):
         tabs = QtWidgets.QTabWidget()
         self.view_tabs = tabs
 
-    def load_view_menu_bar(self):
-        """load menu bar"""
-        bar = QtWidgets.QMenuBar()
-        self.view_menu_bar = bar
-
     def load_view_checkbox_limit_by(self):
         """load limit results by function checkbox"""
         check = QtWidgets.QCheckBox("Limit results to current function")
@@ -319,19 +381,23 @@ class CapaExplorerForm(idaapi.PluginForm):
         analyze_button = QtWidgets.QPushButton("Analyze")
         reset_button = QtWidgets.QPushButton("Reset")
         save_button = QtWidgets.QPushButton("Save")
+        settings_button = QtWidgets.QPushButton("Settings")
 
         analyze_button.clicked.connect(self.slot_analyze)
         reset_button.clicked.connect(self.slot_reset)
         save_button.clicked.connect(self.slot_save)
+        settings_button.clicked.connect(self.slot_settings)
 
         layout = QtWidgets.QHBoxLayout()
         layout.addWidget(analyze_button)
         layout.addWidget(reset_button)
-        layout.addStretch(2)
+        layout.addWidget(settings_button)
+        layout.addStretch(3)
         layout.addWidget(save_button, alignment=QtCore.Qt.AlignRight)
 
         self.view_analyze_button = analyze_button
         self.view_reset_button = reset_button
+        self.view_settings_button = settings_button
         self.view_save_button = save_button
         self.view_buttons = layout
 
@@ -350,7 +416,6 @@ class CapaExplorerForm(idaapi.PluginForm):
         layout.addWidget(self.view_tabs)
         layout.addLayout(self.view_buttons)
         layout.addWidget(self.view_status_label)
-        layout.setMenuBar(self.view_menu_bar)
 
         self.parent.setLayout(layout)
 
@@ -450,27 +515,6 @@ class CapaExplorerForm(idaapi.PluginForm):
 
         self.view_tabs.addTab(tab, "Rule Generator")
 
-    def load_configure_menu(self):
-        """ """
-        actions = (
-            ("Change default rules directory...", "Set default rules directory", self.slot_change_rules_dir),
-            ("Change default rule author...", "Set default rule author", self.slot_change_rule_author),
-            ("Change default rule scope...", "Set default rule scope", self.slot_change_rule_scope),
-        )
-        self.load_menu("Settings", actions)
-
-    def load_menu(self, title, actions):
-        """load menu actions
-
-        @param title: menu name displayed in UI
-        @param actions: tuple of tuples containing action name, tooltip, and slot function
-        """
-        menu = self.view_menu_bar.addMenu(title)
-        for (name, _, slot) in actions:
-            action = QtWidgets.QAction(name, self.parent)
-            action.triggered.connect(slot)
-            menu.addAction(action)
-
     def load_ida_hooks(self):
         """load IDA UI hooks"""
         # map named action (defined in idagui.cfg) to Python function
@@ -567,7 +611,7 @@ class CapaExplorerForm(idaapi.PluginForm):
 
         try:
             # resolve rules directory - check self and settings first, then ask user
-            if not os.path.exists(settings.user.get("rule_path", "")):
+            if not os.path.exists(settings.user.get(CAPA_SETTINGS_RULE_PATH, "")):
                 idaapi.info("Please select a file directory containing capa rules.")
                 path = self.ask_user_directory()
                 if not path:
@@ -575,7 +619,7 @@ class CapaExplorerForm(idaapi.PluginForm):
                         "You must select a file directory containing capa rules before analysis can be run. The standard collection of capa rules can be downloaded from https://github.com/fireeye/capa-rules."
                     )
                     return False
-                settings.user["rule_path"] = path
+                settings.user[CAPA_SETTINGS_RULE_PATH] = path
         except Exception as e:
             logger.error("Failed to load capa rules (error: %s).", e)
             return False
@@ -584,8 +628,9 @@ class CapaExplorerForm(idaapi.PluginForm):
             logger.info("User cancelled analysis.")
             return False
 
-        rule_path = settings.user["rule_path"]
+        rule_path = settings.user[CAPA_SETTINGS_RULE_PATH]
         try:
+            # TODO refactor: this first part is identical to capa.main.get_rules
             if not os.path.exists(rule_path):
                 raise IOError("rule path %s does not exist or cannot be accessed" % rule_path)
 
@@ -601,8 +646,8 @@ class CapaExplorerForm(idaapi.PluginForm):
                         continue
                     for file in files:
                         if not file.endswith(".yml"):
-                            if not (file.endswith(".md") or file.endswith(".git") or file.endswith(".txt")):
-                                # expect to see readme.md, format.md, and maybe a .git directory
+                            if not (file.startswith(".git") or file.endswith((".git", ".md", ".txt"))):
+                                # expect to see .git* files, readme.md, format.md, and maybe a .git directory
                                 # other things maybe are rules, but are mis-named.
                                 logger.warning("skipping non-.yml file: %s", file)
                             continue
@@ -613,7 +658,8 @@ class CapaExplorerForm(idaapi.PluginForm):
             total_paths = len(rule_paths)
             for (i, rule_path) in enumerate(rule_paths):
                 update_wait_box(
-                    "loading capa rules from %s (%d of %d)" % (settings.user["rule_path"], i + 1, total_paths)
+                    "loading capa rules from %s (%d of %d)"
+                    % (settings.user[CAPA_SETTINGS_RULE_PATH], i + 1, total_paths)
                 )
                 if ida_kernwin.user_cancelled():
                     raise UserCancelledError("user cancelled")
@@ -632,12 +678,14 @@ class CapaExplorerForm(idaapi.PluginForm):
             logger.info("User cancelled analysis.")
             return False
         except Exception as e:
-            capa.ida.helpers.inform_user_ida_ui("Failed to load capa rules from %s" % settings.user["rule_path"])
-            logger.error("Failed to load rules from %s (error: %s).", settings.user["rule_path"], e)
+            capa.ida.helpers.inform_user_ida_ui(
+                "Failed to load capa rules from %s" % settings.user[CAPA_SETTINGS_RULE_PATH]
+            )
+            logger.error("Failed to load rules from %s (error: %s).", settings.user[CAPA_SETTINGS_RULE_PATH], e)
             logger.error(
                 "Make sure your file directory contains properly formatted capa rules. You can download the standard collection of capa rules from https://github.com/fireeye/capa-rules."
             )
-            settings.user["rule_path"] = ""
+            settings.user[CAPA_SETTINGS_RULE_PATH] = ""
             return False
 
         self.ruleset_cache = ruleset
@@ -735,7 +783,9 @@ class CapaExplorerForm(idaapi.PluginForm):
             update_wait_box("rendering results")
 
             try:
-                self.doc = capa.render.convert_capabilities_to_result_document(meta, self.ruleset_cache, capabilities)
+                self.doc = capa.render.result_document.convert_capabilities_to_result_document(
+                    meta, self.ruleset_cache, capabilities
+                )
             except Exception as e:
                 logger.error("Failed to render results (error: %s)", e)
                 return False
@@ -743,7 +793,7 @@ class CapaExplorerForm(idaapi.PluginForm):
         try:
             self.model_data.render_capa_doc(self.doc, self.view_show_results_by_function.isChecked())
             self.set_view_status_label(
-                "capa rules directory: %s (%d rules)" % (settings.user["rule_path"], len(self.rules_cache))
+                "capa rules directory: %s (%d rules)" % (settings.user[CAPA_SETTINGS_RULE_PATH], len(self.rules_cache))
             )
         except Exception as e:
             logger.error("Failed to render results (error: %s)", e)
@@ -830,7 +880,7 @@ class CapaExplorerForm(idaapi.PluginForm):
                         if rule.meta.get("capa/subscope-rule"):
                             continue
                         for (ea, _) in res:
-                            func_features[capa.features.MatchedRule(name)].add(ea)
+                            func_features[capa.features.common.MatchedRule(name)].add(ea)
                 except Exception as e:
                     logger.error("Failed to match function/basic block rule scope (error: %s)" % e)
                     return False
@@ -864,7 +914,7 @@ class CapaExplorerForm(idaapi.PluginForm):
                     if rule.meta.get("capa/subscope-rule"):
                         continue
                     for (ea, _) in res:
-                        file_features[capa.features.MatchedRule(name)].add(ea)
+                        file_features[capa.features.common.MatchedRule(name)].add(ea)
             except Exception as e:
                 logger.error("Failed to match file scope rules (error: %s)" % e)
                 return False
@@ -881,14 +931,14 @@ class CapaExplorerForm(idaapi.PluginForm):
             # load preview and feature tree
             self.view_rulegen_preview.load_preview_meta(
                 f.start_ea if f else None,
-                settings.user.get("rulegen_author", "<insert_author>"),
-                settings.user.get("rulegen_scope", "function"),
+                settings.user.get(CAPA_SETTINGS_RULEGEN_AUTHOR, "<insert_author>"),
+                settings.user.get(CAPA_SETTINGS_RULEGEN_SCOPE, "function"),
             )
             self.view_rulegen_features.load_features(file_features, func_features)
 
             # self.view_rulegen_header_label.setText("Function Features (%s)" % trim_function_name(f))
             self.set_view_status_label(
-                "capa rules directory: %s (%d rules)" % (settings.user["rule_path"], len(self.rules_cache))
+                "capa rules directory: %s (%d rules)" % (settings.user[CAPA_SETTINGS_RULE_PATH], len(self.rules_cache))
             )
         except Exception as e:
             logger.error("Failed to render views (error: %s)" % e)
@@ -985,6 +1035,12 @@ class CapaExplorerForm(idaapi.PluginForm):
 
         # create deep copy of current rules, add our new rule
         rules = copy.copy(self.rules_cache)
+
+        # ensure subscope rules are included
+        for sub in rule.extract_subscope_rules():
+            rules.append(sub)
+
+        # include our new rule in the list
         rules.append(rule)
 
         try:
@@ -1066,13 +1122,23 @@ class CapaExplorerForm(idaapi.PluginForm):
         elif self.view_tabs.currentIndex() == 1:
             self.save_function_analysis()
 
+    def slot_settings(self):
+        """ """
+        dialog = CapaSettingsInputDialog("capa explorer settings", parent=self.parent)
+        if dialog.exec_():
+            (
+                settings.user[CAPA_SETTINGS_RULE_PATH],
+                settings.user[CAPA_SETTINGS_RULEGEN_AUTHOR],
+                settings.user[CAPA_SETTINGS_RULEGEN_SCOPE],
+            ) = dialog.get_values()
+
     def save_program_analysis(self):
         """ """
         if not self.doc:
             idaapi.info("No program analysis to save.")
             return
 
-        s = json.dumps(self.doc, sort_keys=True, cls=capa.render.CapaJsonObjectEncoder).encode("utf-8")
+        s = json.dumps(self.doc, sort_keys=True, cls=capa.render.json.CapaJsonObjectEncoder).encode("utf-8")
 
         path = idaapi.ask_file(True, "*.json", "Choose file to save capa program analysis JSON")
         if not path:
@@ -1143,42 +1209,16 @@ class CapaExplorerForm(idaapi.PluginForm):
         """create Qt dialog to ask user for a directory"""
         return str(
             QtWidgets.QFileDialog.getExistingDirectory(
-                self.parent, "Please select a capa rules directory", settings.user["rule_path"]
+                self.parent, "Please select a capa rules directory", settings.user.get(CAPA_SETTINGS_RULE_PATH, "")
             )
         )
 
     def ask_user_capa_rule_file(self):
         """ """
         return QtWidgets.QFileDialog.getSaveFileName(
-            None, "Please select a capa rule to edit", settings.user["rule_path"], "*.yml"
+            None, "Please select a capa rule to edit", settings.user.get(CAPA_SETTINGS_RULE_PATH, ""), "*.yml"
         )[0]
 
-    def slot_change_rule_scope(self):
-        """ """
-        scope = idaapi.ask_str(str(settings.user.get("rulegen_scope", "function")), 0, "Enter default rule scope")
-        if scope:
-            settings.user["rulegen_scope"] = scope
-            idaapi.info("Run analysis again for your changes to take effect.")
-
-    def slot_change_rule_author(self):
-        """ """
-        author = idaapi.ask_str(str(settings.user.get("rulegen_author", "")), 0, "Enter default rule author")
-        if author:
-            settings.user["rulegen_author"] = author
-            idaapi.info("Run analysis again for your changes to take effect.")
-
-    def slot_change_rules_dir(self):
-        """allow user to change rules directory
-
-        user selection stored in settings for future runs
-        """
-        path = self.ask_user_directory()
-        if path:
-            settings.user["rule_path"] = path
-            self.rules_cache = None
-            self.ruleset_cache = None
-            idaapi.info("Run analysis again for your changes to take effect.")
-
     def set_view_status_label(self, text):
         """update status label control
 

capa/ida/plugin/item.py

@@ -6,7 +6,6 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import sys
 import codecs
 
 import idc
@@ -32,7 +31,7 @@ def location_to_hex(location):
     return "%08X" % location
 
 
-class CapaExplorerDataItem(object):
+class CapaExplorerDataItem:
     """store data for CapaExplorerDataModel"""
 
     def __init__(self, parent, data, can_check=True):
@@ -202,7 +201,7 @@ class CapaExplorerRuleMatchItem(CapaExplorerDataItem):
 
     @property
     def source(self):
-        """ return rule contents for display """
+        """return rule contents for display"""
         return self._source
 
 
@@ -328,14 +327,10 @@ class CapaExplorerByteViewItem(CapaExplorerFeatureItem):
         """
         byte_snap = idaapi.get_bytes(location, 32)
 
+        details = ""
         if byte_snap:
             byte_snap = codecs.encode(byte_snap, "hex").upper()
-            if sys.version_info >= (3, 0):
-                details = " ".join([byte_snap[i : i + 2].decode() for i in range(0, len(byte_snap), 2)])
-            else:
-                details = " ".join([byte_snap[i : i + 2] for i in range(0, len(byte_snap), 2)])
-        else:
-            details = ""
+            details = " ".join([byte_snap[i : i + 2].decode() for i in range(0, len(byte_snap), 2)])
 
         super(CapaExplorerByteViewItem, self).__init__(parent, display, location=location, details=details)
         self.ida_highlight = idc.get_color(location, idc.CIC_ITEM)

capa/ida/plugin/model.py

@@ -12,8 +12,10 @@ import idc
 import idaapi
 from PyQt5 import QtGui, QtCore
 
+import capa.rules
 import capa.ida.helpers
 import capa.render.utils as rutils
+import capa.features.common
 from capa.ida.plugin.item import (
     CapaExplorerDataItem,
     CapaExplorerRuleItem,
@@ -488,13 +490,17 @@ class CapaExplorerDataModel(QtCore.QAbstractItemModel):
 
         @param feature: capa feature read from doc
         """
-        if feature[feature["type"]]:
+        key = feature["type"]
+        value = feature[feature["type"]]
+        if value:
+            if key == "string":
+                value = '"%s"' % capa.features.common.escape_string(value)
             if feature.get("description", ""):
-                return "%s(%s = %s)" % (feature["type"], feature[feature["type"]], feature["description"])
+                return "%s(%s = %s)" % (key, value, feature["description"])
             else:
-                return "%s(%s)" % (feature["type"], feature[feature["type"]])
+                return "%s(%s)" % (key, value)
         else:
-            return "%s" % feature["type"]
+            return "%s" % key
 
     def render_capa_doc_feature_node(self, parent, feature, locations, doc):
         """process capa doc feature node
@@ -551,7 +557,14 @@ class CapaExplorerDataModel(QtCore.QAbstractItemModel):
             )
 
         if feature["type"] == "regex":
-            return CapaExplorerStringViewItem(parent, display, location, feature["match"])
+            for s, locations in feature["matches"].items():
+                if location in locations:
+                    return CapaExplorerStringViewItem(
+                        parent, display, location, '"' + capa.features.common.escape_string(s) + '"'
+                    )
+
+            # programming error: the given location should always be found in the regex matches
+            raise ValueError("regex match at location not found")
 
         if feature["type"] == "basicblock":
             return CapaExplorerBlockItem(parent, location)
@@ -576,11 +589,13 @@ class CapaExplorerDataModel(QtCore.QAbstractItemModel):
 
         if feature["type"] in ("string",):
             # display string preview
-            return CapaExplorerStringViewItem(parent, display, location, feature[feature["type"]])
+            return CapaExplorerStringViewItem(
+                parent, display, location, '"%s"' % capa.features.common.escape_string(feature[feature["type"]])
+            )
 
-        if feature["type"] in ("import", "export"):
+        if feature["type"] in ("import", "export", "function-name"):
             # display no preview
-            return CapaExplorerFeatureItem(parent, display=display)
+            return CapaExplorerFeatureItem(parent, location=location, display=display)
 
         raise RuntimeError("unexpected feature type: " + str(feature["type"]))
 

capa/ida/plugin/proxy.py

@@ -5,7 +5,6 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-import six
 from PyQt5 import QtCore
 from PyQt5.QtCore import Qt
 
@@ -208,7 +207,7 @@ class CapaExplorerSearchProxyModel(QtCore.QSortFilterProxyModel):
             if not data:
                 continue
 
-            if not isinstance(data, six.string_types):
+            if not isinstance(data, str):
                 # sanity check: should already be a string, but double check
                 continue
 

capa/ida/plugin/view.py

@@ -14,6 +14,7 @@ from PyQt5 import QtGui, QtCore, QtWidgets
 import capa.rules
 import capa.engine
 import capa.ida.helpers
+import capa.features.common
 import capa.features.basicblock
 from capa.ida.plugin.item import CapaExplorerFunctionItem
 from capa.ida.plugin.model import CapaExplorerDataModel
@@ -178,6 +179,9 @@ def build_context_menu(o, actions):
 
 
 class CapaExplorerRulgenPreview(QtWidgets.QTextEdit):
+
+    INDENT = " " * 2
+
     def __init__(self, parent=None):
         """ """
         super(CapaExplorerRulgenPreview, self).__init__(parent)
@@ -210,12 +214,99 @@ class CapaExplorerRulgenPreview(QtWidgets.QTextEdit):
         self.setText("\n".join(metadata_default))
 
     def keyPressEvent(self, e):
-        """ """
-        if e.key() == QtCore.Qt.Key_Tab:
-            self.insertPlainText(" " * 2)
+        """intercept key press events"""
+        if e.key() in (QtCore.Qt.Key_Tab, QtCore.Qt.Key_Backtab):
+            # apparently it's not easy to implement tabs as spaces, or multi-line tab or SHIFT + Tab
+            # so we need to implement it ourselves so we can retain properly formatted capa rules
+            # when a user uses the Tab key
+            if self.textCursor().selection().isEmpty():
+                # single line, only worry about Tab
+                if e.key() == QtCore.Qt.Key_Tab:
+                    self.insertPlainText(self.INDENT)
+            else:
+                # multi-line tab or SHIFT + Tab
+                cur = self.textCursor()
+                select_start_ppos = cur.selectionStart()
+                select_end_ppos = cur.selectionEnd()
+
+                scroll_ppos = self.verticalScrollBar().sliderPosition()
+
+                # determine lineno for first selected line, and column
+                cur.setPosition(select_start_ppos)
+                start_lineno = self.count_previous_lines_from_block(cur.block())
+                start_lineco = cur.columnNumber()
+
+                # determine lineno for last selected line
+                cur.setPosition(select_end_ppos)
+                end_lineno = self.count_previous_lines_from_block(cur.block())
+
+                # now we need to indent or dedent the selected lines. for now, we read the text, modify
+                # the lines between start_lineno and end_lineno accordingly, and then reset the view
+                # this might not be the best solution, but it avoids messing around with cursor positions
+                # to determine the beginning of lines
+
+                plain = self.toPlainText().splitlines()
+
+                if e.key() == QtCore.Qt.Key_Tab:
+                    # user Tab, indent selected lines
+                    lines_modified = end_lineno - start_lineno
+                    first_modified = True
+                    change = [self.INDENT + line for line in plain[start_lineno : end_lineno + 1]]
+                else:
+                    # user SHIFT + Tab, dedent selected lines
+                    lines_modified = 0
+                    first_modified = False
+                    change = []
+                    for (lineno, line) in enumerate(plain[start_lineno : end_lineno + 1]):
+                        if line.startswith(self.INDENT):
+                            if lineno == 0:
+                                # keep track if first line is modified, so we can properly display
+                                # the text selection later
+                                first_modified = True
+                            lines_modified += 1
+                            line = line[len(self.INDENT) :]
+                        change.append(line)
+
+                # apply modifications, and reset view
+                plain[start_lineno : end_lineno + 1] = change
+                self.setPlainText("\n".join(plain) + "\n")
+
+                # now we need to properly adjust the selection positions, so users don't have to
+                # re-select when indenting or dedenting the same lines repeatedly
+                if e.key() == QtCore.Qt.Key_Tab:
+                    # user Tab, increase increment selection positions
+                    select_start_ppos += len(self.INDENT)
+                    select_end_ppos += (lines_modified * len(self.INDENT)) + len(self.INDENT)
+                elif lines_modified:
+                    # user SHIFT + Tab, decrease selection positions
+                    if start_lineco not in (0, 1) and first_modified:
+                        # only decrease start position if not in first column
+                        select_start_ppos -= len(self.INDENT)
+                    select_end_ppos -= lines_modified * len(self.INDENT)
+
+                # apply updated selection and restore previous scroll position
+                self.set_selection(select_start_ppos, select_end_ppos, len(self.toPlainText()))
+                self.verticalScrollBar().setSliderPosition(scroll_ppos)
         else:
             super(CapaExplorerRulgenPreview, self).keyPressEvent(e)
 
+    def count_previous_lines_from_block(self, block):
+        """calculate number of lines preceding block"""
+        count = 0
+        while True:
+            block = block.previous()
+            if not block.isValid():
+                break
+            count += block.lineCount()
+        return count
+
+    def set_selection(self, start, end, max):
+        """set text selection"""
+        cursor = self.textCursor()
+        cursor.setPosition(start)
+        cursor.setPosition(end if end < max else max, QtGui.QTextCursor.KeepAnchor)
+        self.setTextCursor(cursor)
+
 
 class CapaExplorerRulgenEditor(QtWidgets.QTreeWidget):
 
@@ -325,6 +416,11 @@ class CapaExplorerRulgenEditor(QtWidgets.QTreeWidget):
         # create a new parent under root node, by default; new node added last position in tree
         new_parent = self.new_expression_node(self.root, (action.data()[0], ""))
 
+        if "basic block" in action.data()[0]:
+            # add default child expression when nesting under basic block
+            new_parent.setExpanded(True)
+            new_parent = self.new_expression_node(new_parent, ("- or:", ""))
+
         for o in self.get_features(selected=True):
             # take child from its parent by index, add to new parent
             new_parent.addChild(o.parent().takeChild(o.parent().indexOfChild(o)))
@@ -335,6 +431,15 @@ class CapaExplorerRulgenEditor(QtWidgets.QTreeWidget):
     def slot_edit_expression(self, action):
         """ """
         expression, o = action.data()
+        if "basic block" in expression and "basic block" not in o.text(
+            CapaExplorerRulgenEditor.get_column_feature_index()
+        ):
+            # current expression is "basic block", and not changing to "basic block" expression
+            children = o.takeChildren()
+            new_parent = self.new_expression_node(o, ("- or:", ""))
+            for child in children:
+                new_parent.addChild(child)
+            new_parent.setExpanded(True)
         o.setText(CapaExplorerRulgenEditor.get_column_feature_index(), expression)
 
     def slot_clear_all(self, action):
@@ -520,11 +625,23 @@ class CapaExplorerRulgenEditor(QtWidgets.QTreeWidget):
 
         # single features
         for (k, v) in filter(lambda t: t[1] == 1, counted):
-            self.new_feature_node(self.root, ("- %s: %s" % (k.name.lower(), k.get_value_str()), ""))
+            if isinstance(k, (capa.features.common.String,)):
+                value = '"%s"' % capa.features.common.escape_string(k.get_value_str())
+            else:
+                value = k.get_value_str()
+            self.new_feature_node(self.root, ("- %s: %s" % (k.name.lower(), value), ""))
 
         # n > 1 features
         for (k, v) in filter(lambda t: t[1] > 1, counted):
-            self.new_feature_node(self.root, ("- count(%s): %d" % (str(k), v), ""))
+            if k.value:
+                if isinstance(k, (capa.features.common.String,)):
+                    value = '"%s"' % capa.features.common.escape_string(k.get_value_str())
+                else:
+                    value = k.get_value_str()
+                display = "- count(%s(%s)): %d" % (k.name.lower(), value, v)
+            else:
+                display = "- count(%s): %d" % (k.name.lower(), v)
+            self.new_feature_node(self.root, (display, ""))
 
         self.expandAll()
         self.update_preview()
@@ -699,9 +816,11 @@ class CapaExplorerRulegenFeatures(QtWidgets.QTreeWidget):
         if text:
             for o in iterate_tree(self):
                 data = o.data(0, 0x100)
-                if data and text.lower() not in data.get_value_str().lower():
-                    o.setHidden(True)
-                    continue
+                if data:
+                    to_match = data.get_value_str()
+                    if not to_match or text.lower() not in to_match.lower():
+                        o.setHidden(True)
+                        continue
                 o.setHidden(False)
                 o.setExpanded(True)
         else:
@@ -776,16 +895,20 @@ class CapaExplorerRulegenFeatures(QtWidgets.QTreeWidget):
         def format_address(e):
             return "%X" % e if e else ""
 
+        def format_feature(feature):
+            """ """
+            name = feature.name.lower()
+            value = feature.get_value_str()
+            if isinstance(feature, (capa.features.common.String,)):
+                value = '"%s"' % capa.features.common.escape_string(value)
+            return "%s(%s)" % (name, value)
+
         for (feature, eas) in sorted(features.items(), key=lambda k: sorted(k[1])):
             if isinstance(feature, capa.features.basicblock.BasicBlock):
                 # filter basic blocks for now, we may want to add these back in some time
                 # in the future
                 continue
 
-            if isinstance(feature, capa.features.String):
-                # strip string for display
-                feature.value = feature.value.strip()
-
             # level 0
             if type(feature) not in self.parent_items:
                 self.parent_items[type(feature)] = self.new_parent_node(parent, (feature.name.lower(),))
@@ -794,20 +917,22 @@ class CapaExplorerRulegenFeatures(QtWidgets.QTreeWidget):
             if feature not in self.parent_items:
                 if len(eas) > 1:
                     self.parent_items[feature] = self.new_parent_node(
-                        self.parent_items[type(feature)], (str(feature),), feature=feature
+                        self.parent_items[type(feature)], (format_feature(feature),), feature=feature
                     )
                 else:
                     self.parent_items[feature] = self.new_leaf_node(
-                        self.parent_items[type(feature)], (str(feature),), feature=feature
+                        self.parent_items[type(feature)], (format_feature(feature),), feature=feature
                     )
 
             # level n > 1
             if len(eas) > 1:
                 for ea in sorted(eas):
-                    self.new_leaf_node(self.parent_items[feature], (str(feature), format_address(ea)), feature=feature)
+                    self.new_leaf_node(
+                        self.parent_items[feature], (format_feature(feature), format_address(ea)), feature=feature
+                    )
             else:
                 ea = eas.pop()
-                for (i, v) in enumerate((str(feature), format_address(ea))):
+                for (i, v) in enumerate((format_feature(feature), format_address(ea))):
                     self.parent_items[feature].setText(i, v)
                 self.parent_items[feature].setData(0, 0x100, feature)
 

capa/main.py

@@ -1,4 +1,4 @@
-#!/usr/bin/env python2
+#!/usr/bin/env python3
 """
 Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
 Licensed under the Apache License, Version 2.0 (the "License");
@@ -10,13 +10,18 @@ See the License for the specific language governing permissions and limitations
 """
 import os
 import sys
+import gzip
+import time
 import hashlib
 import logging
 import os.path
 import argparse
 import datetime
 import textwrap
+import itertools
+import contextlib
 import collections
+from typing import Any, Dict, List, Tuple
 
 import halo
 import tqdm
@@ -24,22 +29,40 @@ import colorama
 
 import capa.rules
 import capa.engine
-import capa.render
 import capa.version
-import capa.features
+import capa.render.json
+import capa.render.default
+import capa.render.verbose
+import capa.features.common
 import capa.features.freeze
+import capa.render.vverbose
 import capa.features.extractors
-from capa.helpers import oint, get_file_taste
+import capa.features.extractors.pefile
+from capa.rules import Rule, RuleSet
+from capa.engine import FeatureSet, MatchResults
+from capa.helpers import get_file_taste
+from capa.features.extractors.base_extractor import FunctionHandle, FeatureExtractor
 
 RULES_PATH_DEFAULT_STRING = "(embedded rules)"
+SIGNATURES_PATH_DEFAULT_STRING = "(embedded signatures)"
 SUPPORTED_FILE_MAGIC = set([b"MZ"])
 BACKEND_VIV = "vivisect"
 BACKEND_SMDA = "smda"
+EXTENSIONS_SHELLCODE_32 = ("sc32", "raw32")
+EXTENSIONS_SHELLCODE_64 = ("sc64", "raw64")
 
 
 logger = logging.getLogger("capa")
 
 
+@contextlib.contextmanager
+def timing(msg: str):
+    t0 = time.time()
+    yield
+    t1 = time.time()
+    logger.debug("perf: %s: %0.2fs", msg, t1 - t0)
+
+
 def set_vivisect_log_level(level):
     logging.getLogger("vivisect").setLevel(level)
     logging.getLogger("vivisect.base").setLevel(level)
@@ -49,12 +72,12 @@ def set_vivisect_log_level(level):
     logging.getLogger("envi.codeflow").setLevel(level)
 
 
-def find_function_capabilities(ruleset, extractor, f):
+def find_function_capabilities(ruleset: RuleSet, extractor: FeatureExtractor, f: FunctionHandle):
     # contains features from:
     #  - insns
     #  - function
-    function_features = collections.defaultdict(set)
-    bb_matches = collections.defaultdict(list)
+    function_features = collections.defaultdict(set)  # type: FeatureSet
+    bb_matches = collections.defaultdict(list)  # type: MatchResults
 
     for feature, va in extractor.extract_function_features(f):
         function_features[feature].add(va)
@@ -74,19 +97,19 @@ def find_function_capabilities(ruleset, extractor, f):
                 bb_features[feature].add(va)
                 function_features[feature].add(va)
 
-        _, matches = capa.engine.match(ruleset.basic_block_rules, bb_features, oint(bb))
+        _, matches = capa.engine.match(ruleset.basic_block_rules, bb_features, int(bb))
 
         for rule_name, res in matches.items():
             bb_matches[rule_name].extend(res)
             for va, _ in res:
-                function_features[capa.features.MatchedRule(rule_name)].add(va)
+                function_features[capa.features.common.MatchedRule(rule_name)].add(va)
 
-    _, function_matches = capa.engine.match(ruleset.function_rules, function_features, oint(f))
+    _, function_matches = capa.engine.match(ruleset.function_rules, function_features, int(f))
     return function_matches, bb_matches, len(function_features)
 
 
-def find_file_capabilities(ruleset, extractor, function_features):
-    file_features = collections.defaultdict(set)
+def find_file_capabilities(ruleset: RuleSet, extractor: FeatureExtractor, function_features: FeatureSet):
+    file_features = collections.defaultdict(set)  # type: FeatureSet
 
     for feature, va in extractor.extract_file_features():
         # not all file features may have virtual addresses.
@@ -106,16 +129,17 @@ def find_file_capabilities(ruleset, extractor, function_features):
     return matches, len(file_features)
 
 
-def find_capabilities(ruleset, extractor, disable_progress=None):
-    all_function_matches = collections.defaultdict(list)
-    all_bb_matches = collections.defaultdict(list)
+def find_capabilities(ruleset: RuleSet, extractor: FeatureExtractor, disable_progress=None) -> Tuple[MatchResults, Any]:
+    all_function_matches = collections.defaultdict(list)  # type: MatchResults
+    all_bb_matches = collections.defaultdict(list)  # type: MatchResults
 
     meta = {
         "feature_counts": {
             "file": 0,
             "functions": {},
-        }
-    }
+        },
+        "library_functions": {},
+    }  # type: Dict[str, Any]
 
     pbar = tqdm.tqdm
     if disable_progress:
@@ -123,30 +147,53 @@ def find_capabilities(ruleset, extractor, disable_progress=None):
         # to disable progress completely
         pbar = lambda s, *args, **kwargs: s
 
-    for f in pbar(list(extractor.get_functions()), desc="matching", unit=" functions"):
+    functions = list(extractor.get_functions())
+    n_funcs = len(functions)
+
+    pb = pbar(functions, desc="matching", unit=" functions", postfix="skipped 0 library functions")
+    for f in pb:
+        function_address = int(f)
+
+        if extractor.is_library_function(function_address):
+            function_name = extractor.get_function_name(function_address)
+            logger.debug("skipping library function 0x%x (%s)", function_address, function_name)
+            meta["library_functions"][function_address] = function_name
+            n_libs = len(meta["library_functions"])
+            percentage = 100 * (n_libs / n_funcs)
+            if isinstance(pb, tqdm.tqdm):
+                pb.set_postfix_str("skipped %d library functions (%d%%)" % (n_libs, percentage))
+            continue
+
         function_matches, bb_matches, feature_count = find_function_capabilities(ruleset, extractor, f)
-        meta["feature_counts"]["functions"][f.__int__()] = feature_count
-        logger.debug("analyzed function 0x%x and extracted %d features", f.__int__(), feature_count)
+        meta["feature_counts"]["functions"][function_address] = feature_count
+        logger.debug("analyzed function 0x%x and extracted %d features", function_address, feature_count)
 
         for rule_name, res in function_matches.items():
             all_function_matches[rule_name].extend(res)
         for rule_name, res in bb_matches.items():
             all_bb_matches[rule_name].extend(res)
 
-    # mapping from matched rule feature to set of addresses at which it matched.
-    # schema: Dic[MatchedRule: Set[int]
-    function_features = {
-        capa.features.MatchedRule(rule_name): set(map(lambda p: p[0], results))
-        for rule_name, results in all_function_matches.items()
-    }
+    # collection of features that captures the rule matches within function and BB scopes.
+    # mapping from feature (matched rule) to set of addresses at which it matched.
+    function_and_lower_features = {
+        capa.features.common.MatchedRule(rule_name): set(map(lambda p: p[0], results))
+        for rule_name, results in itertools.chain(all_function_matches.items(), all_bb_matches.items())
+    }  # type: FeatureSet
 
-    all_file_matches, feature_count = find_file_capabilities(ruleset, extractor, function_features)
+    all_file_matches, feature_count = find_file_capabilities(ruleset, extractor, function_and_lower_features)
     meta["feature_counts"]["file"] = feature_count
 
-    matches = {}
-    matches.update(all_bb_matches)
-    matches.update(all_function_matches)
-    matches.update(all_file_matches)
+    matches = {
+        rule_name: results
+        for rule_name, results in itertools.chain(
+            # each rule exists in exactly one scope,
+            # so there won't be any overlap among these following MatchResults,
+            # and we can merge the dictionaries naively.
+            all_bb_matches.items(),
+            all_function_matches.items(),
+            all_file_matches.items(),
+        )
+    }
 
     return matches, meta
 
@@ -158,61 +205,37 @@ def has_rule_with_namespace(rules, capabilities, rule_cat):
     return False
 
 
-def has_file_limitation(rules, capabilities, is_standalone=True):
-    file_limitations = {
-        # capa will likely detect installer specific functionality.
-        # this is probably not what the user wants.
-        "executable/installer": [
-            " This sample appears to be an installer.",
-            " ",
-            " capa cannot handle installers well. This means the results may be misleading or incomplete."
-            " You should try to understand the install mechanism and analyze created files with capa.",
-        ],
-        # capa won't detect much in .NET samples.
-        # it might match some file-level things.
-        # for consistency, bail on things that we don't support.
-        "runtime/dotnet": [
-            " This sample appears to be a .NET module.",
-            " ",
-            " .NET is a cross-platform framework for running managed applications.",
-            " capa cannot handle non-native files. This means that the results may be misleading or incomplete.",
-            " You may have to analyze the file manually, using a tool like the .NET decompiler dnSpy.",
-        ],
-        # capa will detect dozens of capabilities for AutoIt samples,
-        # but these are due to the AutoIt runtime, not the payload script.
-        # so, don't confuse the user with FP matches - bail instead
-        "compiler/autoit": [
-            " This sample appears to be compiled with AutoIt.",
-            " ",
-            " AutoIt is a freeware BASIC-like scripting language designed for automating the Windows GUI.",
-            " capa cannot handle AutoIt scripts. This means that the results will be misleading or incomplete.",
-            " You may have to analyze the file manually, using a tool like the AutoIt decompiler MyAut2Exe.",
-        ],
-        # capa won't detect much in packed samples
-        "anti-analysis/packer/": [
-            " This sample appears to be packed.",
-            " ",
-            " Packed samples have often been obfuscated to hide their logic.",
-            " capa cannot handle obfuscation well. This means the results may be misleading or incomplete.",
-            " If possible, you should try to unpack this input file before analyzing it with capa.",
-        ],
-    }
+def is_internal_rule(rule: Rule) -> bool:
+    return rule.meta.get("namespace", "").startswith("internal/")
 
-    for category, dialogue in file_limitations.items():
-        if not has_rule_with_namespace(rules, capabilities, category):
+
+def is_file_limitation_rule(rule: Rule) -> bool:
+    return rule.meta.get("namespace", "") == "internal/limitation/file"
+
+
+def has_file_limitation(rules: RuleSet, capabilities: MatchResults, is_standalone=True) -> bool:
+    file_limitation_rules = list(filter(is_file_limitation_rule, rules.rules.values()))
+
+    for file_limitation_rule in file_limitation_rules:
+        if file_limitation_rule.name not in capabilities:
             continue
+
         logger.warning("-" * 80)
-        for line in dialogue:
-            logger.warning(line)
+        for line in file_limitation_rule.meta.get("description", "").split("\n"):
+            logger.warning(" " + line)
+        logger.warning(" Identified via rule: %s", file_limitation_rule.name)
         if is_standalone:
             logger.warning(" ")
             logger.warning(" Use -v or -vv if you really want to see the capabilities identified by capa.")
         logger.warning("-" * 80)
+
+        # bail on first file limitation
         return True
+
     return False
 
 
-def is_supported_file_type(sample):
+def is_supported_file_type(sample: str) -> bool:
     """
     Return if this is a supported file based on magic header values
     """
@@ -227,26 +250,31 @@ def is_supported_file_type(sample):
 SHELLCODE_BASE = 0x690000
 
 
-def get_shellcode_vw(sample, arch="auto", should_save=True):
+def get_shellcode_vw(sample, arch="auto"):
     """
-    Return shellcode workspace using explicit arch or via auto detect
+    Return shellcode workspace using explicit arch or via auto detect.
+    The workspace is *not* analyzed nor saved. Its up to the caller to do this.
+    Then, they can register FLIRT analyzers or decide not to write to disk.
     """
     import viv_utils
 
     with open(sample, "rb") as f:
         sample_bytes = f.read()
+
     if arch == "auto":
         # choose arch with most functions, idea by Jay G.
         vw_cands = []
         for arch in ["i386", "amd64"]:
             vw_cands.append(
-                viv_utils.getShellcodeWorkspace(sample_bytes, arch, base=SHELLCODE_BASE, should_save=should_save)
+                viv_utils.getShellcodeWorkspace(
+                    sample_bytes, arch, base=SHELLCODE_BASE, analyze=False, should_save=False
+                )
             )
         if not vw_cands:
             raise ValueError("could not generate vivisect workspace")
         vw = max(vw_cands, key=lambda vw: len(vw.getFunctions()))
     else:
-        vw = viv_utils.getShellcodeWorkspace(sample_bytes, arch, base=SHELLCODE_BASE, should_save=should_save)
+        vw = viv_utils.getShellcodeWorkspace(sample_bytes, arch, base=SHELLCODE_BASE, analyze=False, should_save=False)
 
     vw.setMeta("StorageName", "%s.viv" % sample)
 
@@ -264,55 +292,162 @@ def get_meta_str(vw):
     return "%s, number of functions: %d" % (", ".join(meta), len(vw.getFunctions()))
 
 
+def load_flirt_signature(path):
+    # lazy import enables us to only require flirt here and not in IDA, for example
+    import flirt
+
+    if path.endswith(".sig"):
+        with open(path, "rb") as f:
+            with timing("flirt: parsing .sig: " + path):
+                sigs = flirt.parse_sig(f.read())
+
+    elif path.endswith(".pat"):
+        with open(path, "rb") as f:
+            with timing("flirt: parsing .pat: " + path):
+                sigs = flirt.parse_pat(f.read().decode("utf-8").replace("\r\n", "\n"))
+
+    elif path.endswith(".pat.gz"):
+        with gzip.open(path, "rb") as f:
+            with timing("flirt: parsing .pat.gz: " + path):
+                sigs = flirt.parse_pat(f.read().decode("utf-8").replace("\r\n", "\n"))
+
+    else:
+        raise ValueError("unexpect signature file extension: " + path)
+
+    return sigs
+
+
+def register_flirt_signature_analyzers(vw, sigpaths):
+    """
+    args:
+      vw (vivisect.VivWorkspace):
+      sigpaths (List[str]): file system paths of .sig/.pat files
+    """
+    # lazy import enables us to only require flirt here and not in IDA, for example
+    import flirt
+    import viv_utils.flirt
+
+    for sigpath in sigpaths:
+        try:
+            sigs = load_flirt_signature(sigpath)
+        except ValueError as e:
+            logger.warning("could not load %s: %s", sigpath, str(e))
+            continue
+
+        logger.debug("flirt: sig count: %d", len(sigs))
+
+        with timing("flirt: compiling sigs"):
+            matcher = flirt.compile(sigs)
+
+        analyzer = viv_utils.flirt.FlirtFunctionAnalyzer(matcher, sigpath)
+        logger.debug("registering viv function analyzer: %s", repr(analyzer))
+        viv_utils.flirt.addFlirtFunctionAnalyzer(vw, analyzer)
+
+
+def is_running_standalone() -> bool:
+    """
+    are we running from a PyInstaller'd executable?
+    if so, then we'll be able to access `sys._MEIPASS` for the packaged resources.
+    """
+    return hasattr(sys, "frozen") and hasattr(sys, "_MEIPASS")
+
+
+def get_default_root() -> str:
+    """
+    get the file system path to the default resources directory.
+    under PyInstaller, this comes from _MEIPASS.
+    under source, this is the root directory of the project.
+    """
+    if is_running_standalone():
+        # pylance/mypy don't like `sys._MEIPASS` because this isn't standard.
+        # its injected by pyinstaller.
+        # so we'll fetch this attribute dynamically.
+        return getattr(sys, "_MEIPASS")
+    else:
+        return os.path.join(os.path.dirname(__file__), "..")
+
+
+def get_default_signatures() -> List[str]:
+    """
+    compute a list of file system paths to the default FLIRT signatures.
+    """
+    sigs_path = os.path.join(get_default_root(), "sigs")
+    logger.debug("signatures path: %s", sigs_path)
+
+    ret = []
+    for root, dirs, files in os.walk(sigs_path):
+        for file in files:
+            if not (file.endswith(".pat") or file.endswith(".pat.gz") or file.endswith(".sig")):
+                continue
+
+            ret.append(os.path.join(root, file))
+
+    return ret
+
+
 class UnsupportedFormatError(ValueError):
     pass
 
 
-def get_workspace(path, format, should_save=True):
+def get_workspace(path, format, sigpaths):
+    """
+    load the program at the given path into a vivisect workspace using the given format.
+    also apply the given FLIRT signatures.
+
+    supported formats:
+      - pe
+      - sc32
+      - sc64
+      - auto
+
+    this creates and analyzes the workspace; however, it does *not* save the workspace.
+    this is the responsibility of the caller.
+    """
+
+    # lazy import enables us to not require viv if user wants SMDA, for example.
     import viv_utils
 
     logger.debug("generating vivisect workspace for: %s", path)
     if format == "auto":
         if not is_supported_file_type(path):
             raise UnsupportedFormatError()
-        vw = viv_utils.getWorkspace(path, should_save=should_save)
+
+        # don't analyze, so that we can add our Flirt function analyzer first.
+        vw = viv_utils.getWorkspace(path, analyze=False, should_save=False)
     elif format == "pe":
-        vw = viv_utils.getWorkspace(path, should_save=should_save)
+        vw = viv_utils.getWorkspace(path, analyze=False, should_save=False)
     elif format == "sc32":
-        vw = get_shellcode_vw(path, arch="i386", should_save=should_save)
+        # these are not analyzed nor saved.
+        vw = get_shellcode_vw(path, arch="i386")
     elif format == "sc64":
-        vw = get_shellcode_vw(path, arch="amd64", should_save=should_save)
+        vw = get_shellcode_vw(path, arch="amd64")
     else:
         raise ValueError("unexpected format: " + format)
+
+    register_flirt_signature_analyzers(vw, sigpaths)
+
+    vw.analyze()
+
     logger.debug("%s", get_meta_str(vw))
     return vw
 
 
-def get_extractor_py2(path, format, disable_progress=False):
-    import capa.features.extractors.viv
-
-    with halo.Halo(text="analyzing program", spinner="simpleDots", stream=sys.stderr, enabled=not disable_progress):
-        vw = get_workspace(path, format, should_save=False)
-
-        try:
-            vw.saveWorkspace()
-        except IOError:
-            # see #168 for discussion around how to handle non-writable directories
-            logger.info("source directory is not writable, won't save intermediate workspace")
-
-    return capa.features.extractors.viv.VivisectFeatureExtractor(vw, path)
-
-
 class UnsupportedRuntimeError(RuntimeError):
     pass
 
 
-def get_extractor_py3(path, format, backend, disable_progress=False):
+def get_extractor(
+    path: str, format: str, backend: str, sigpaths: List[str], should_save_workspace=False, disable_progress=False
+) -> FeatureExtractor:
+    """
+    raises:
+      UnsupportedFormatError:
+    """
     if backend == "smda":
         from smda.SmdaConfig import SmdaConfig
         from smda.Disassembler import Disassembler
 
-        import capa.features.extractors.smda
+        import capa.features.extractors.smda.extractor
 
         smda_report = None
         with halo.Halo(text="analyzing program", spinner="simpleDots", stream=sys.stderr, enabled=not disable_progress):
@@ -321,34 +456,31 @@ def get_extractor_py3(path, format, backend, disable_progress=False):
             smda_disasm = Disassembler(config)
             smda_report = smda_disasm.disassembleFile(path)
 
-        return capa.features.extractors.smda.SmdaFeatureExtractor(smda_report, path)
+        return capa.features.extractors.smda.extractor.SmdaFeatureExtractor(smda_report, path)
     else:
-        import capa.features.extractors.viv
+        import capa.features.extractors.viv.extractor
 
         with halo.Halo(text="analyzing program", spinner="simpleDots", stream=sys.stderr, enabled=not disable_progress):
-            vw = get_workspace(path, format, should_save=False)
+            if format == "auto" and path.endswith(EXTENSIONS_SHELLCODE_32):
+                format = "sc32"
+            elif format == "auto" and path.endswith(EXTENSIONS_SHELLCODE_64):
+                format = "sc64"
+            vw = get_workspace(path, format, sigpaths)
 
-            try:
-                vw.saveWorkspace()
-            except IOError:
-                # see #168 for discussion around how to handle non-writable directories
-                logger.info("source directory is not writable, won't save intermediate workspace")
+            if should_save_workspace:
+                logger.debug("saving workspace")
+                try:
+                    vw.saveWorkspace()
+                except IOError:
+                    # see #168 for discussion around how to handle non-writable directories
+                    logger.info("source directory is not writable, won't save intermediate workspace")
+            else:
+                logger.debug("CAPA_SAVE_WORKSPACE unset, not saving workspace")
 
-        return capa.features.extractors.viv.VivisectFeatureExtractor(vw, path)
+        return capa.features.extractors.viv.extractor.VivisectFeatureExtractor(vw, path)
 
 
-def get_extractor(path, format, backend, disable_progress=False):
-    """
-    raises:
-      UnsupportedFormatError:
-    """
-    if sys.version_info >= (3, 0):
-        return get_extractor_py3(path, format, backend, disable_progress=disable_progress)
-    else:
-        return get_extractor_py2(path, format, disable_progress=disable_progress)
-
-
-def is_nursery_rule_path(path):
+def is_nursery_rule_path(path: str) -> bool:
     """
     The nursery is a spot for rules that have not yet been fully polished.
     For example, they may not have references to public example of a technique.
@@ -361,7 +493,7 @@ def is_nursery_rule_path(path):
     return "nursery" in path
 
 
-def get_rules(rule_path, disable_progress=False):
+def get_rules(rule_path: str, disable_progress=False) -> List[Rule]:
     if not os.path.exists(rule_path):
         raise IOError("rule path %s does not exist or cannot be accessed" % rule_path)
 
@@ -379,8 +511,8 @@ def get_rules(rule_path, disable_progress=False):
 
             for file in files:
                 if not file.endswith(".yml"):
-                    if not (file.endswith(".md") or file.endswith(".git") or file.endswith(".txt")):
-                        # expect to see readme.md, format.md, and maybe a .git directory
+                    if not (file.startswith(".git") or file.endswith((".git", ".md", ".txt"))):
+                        # expect to see .git* files, readme.md, format.md, and maybe a .git directory
                         # other things maybe are rules, but are mis-named.
                         logger.warning("skipping non-.yml file: %s", file)
                     continue
@@ -388,7 +520,7 @@ def get_rules(rule_path, disable_progress=False):
                 rule_path = os.path.join(root, file)
                 rule_paths.append(rule_path)
 
-    rules = []
+    rules = []  # type: List[Rule]
 
     pbar = tqdm.tqdm
     if disable_progress:
@@ -396,7 +528,7 @@ def get_rules(rule_path, disable_progress=False):
         # to disable progress completely
         pbar = lambda s, *args, **kwargs: s
 
-    for rule_path in pbar(list(rule_paths), desc="loading ", unit="     rules"):
+    for rule_path in pbar(list(rule_paths), desc="loading ", unit=" rules"):
         try:
             rule = capa.rules.Rule.from_yaml_file(rule_path)
         except capa.rules.InvalidRule:
@@ -412,6 +544,34 @@ def get_rules(rule_path, disable_progress=False):
     return rules
 
 
+def get_signatures(sigs_path):
+    if not os.path.exists(sigs_path):
+        raise IOError("signatures path %s does not exist or cannot be accessed" % sigs_path)
+
+    paths = []
+    if os.path.isfile(sigs_path):
+        paths.append(sigs_path)
+    elif os.path.isdir(sigs_path):
+        logger.debug("reading signatures from directory %s", os.path.abspath(os.path.normpath(sigs_path)))
+        for root, dirs, files in os.walk(sigs_path):
+            for file in files:
+                if file.endswith((".pat", ".pat.gz", ".sig")):
+                    sig_path = os.path.join(root, file)
+                    paths.append(sig_path)
+
+    # nicely normalize and format path so that debugging messages are clearer
+    paths = [os.path.abspath(os.path.normpath(path)) for path in paths]
+
+    # load signatures in deterministic order: the alphabetic sorting of filename.
+    # this means that `0_sigs.pat` loads before `1_sigs.pat`.
+    paths = sorted(paths, key=os.path.basename)
+
+    for path in paths:
+        logger.debug("found signature file: %s", path)
+
+    return paths
+
+
 def collect_metadata(argv, sample_path, rules_path, format, extractor):
     md5 = hashlib.md5()
     sha1 = hashlib.sha1()
@@ -460,7 +620,7 @@ def install_common_args(parser, wanted=None):
       wanted (Set[str]): collection of arguments to opt-into, including:
         - "sample": required positional argument to input file.
         - "format": flag to override file format.
-        - "backend": flag to override analysis backend under py3.
+        - "backend": flag to override analysis backend.
         - "rules": flag to override path to capa rules.
         - "tag": flag to override/specify which rules to match.
     """
@@ -498,22 +658,11 @@ def install_common_args(parser, wanted=None):
     #
 
     if "sample" in wanted:
-        if sys.version_info >= (3, 0):
-            parser.add_argument(
-                # Python 3 str handles non-ASCII arguments correctly
-                "sample",
-                type=str,
-                help="path to sample to analyze",
-            )
-        else:
-            parser.add_argument(
-                # in #328 we noticed that the sample path is not handled correctly if it contains non-ASCII characters
-                # https://stackoverflow.com/a/22947334/ offers a solution and decoding using getfilesystemencoding works
-                # in our testing, however other sources suggest `sys.stdin.encoding` (https://stackoverflow.com/q/4012571/)
-                "sample",
-                type=lambda s: s.decode(sys.getfilesystemencoding()),
-                help="path to sample to analyze",
-            )
+        parser.add_argument(
+            "sample",
+            type=str,
+            help="path to sample to analyze",
+        )
 
     if "format" in wanted:
         formats = [
@@ -532,15 +681,15 @@ def install_common_args(parser, wanted=None):
             help="select sample format, %s" % format_help,
         )
 
-    if "backend" in wanted and sys.version_info >= (3, 0):
-        parser.add_argument(
-            "-b",
-            "--backend",
-            type=str,
-            help="select the backend to use",
-            choices=(BACKEND_VIV, BACKEND_SMDA),
-            default=BACKEND_VIV,
-        )
+        if "backend" in wanted:
+            parser.add_argument(
+                "-b",
+                "--backend",
+                type=str,
+                help="select the backend to use",
+                choices=(BACKEND_VIV, BACKEND_SMDA),
+                default=BACKEND_VIV,
+            )
 
     if "rules" in wanted:
         parser.add_argument(
@@ -551,6 +700,15 @@ def install_common_args(parser, wanted=None):
             help="path to rule file or directory, use embedded rules by default",
         )
 
+    if "signatures" in wanted:
+        parser.add_argument(
+            "-s",
+            "--signatures",
+            type=str,
+            default=SIGNATURES_PATH_DEFAULT_STRING,
+            help="path to .sig/.pat file or directory used to identify library functions, use embedded signatures by default",
+        )
+
     if "tag" in wanted:
         parser.add_argument("-t", "--tag", type=str, help="filter on rule meta field values")
 
@@ -559,6 +717,9 @@ def handle_common_args(args):
     """
     handle the global config specified by `install_common_args`,
     such as configuring logging/coloring/etc.
+    the following fields will be overwritten when present:
+      - rules: file system path to rule files.
+      - signatures: file system path to signature files.
 
     args:
       args (argparse.Namespace): parsed arguments that included at least `install_common_args` args.
@@ -576,10 +737,9 @@ def handle_common_args(args):
     # disable vivisect-related logging, it's verbose and not relevant for capa users
     set_vivisect_log_level(logging.CRITICAL)
 
-    # py2 doesn't know about cp65001, which is a variant of utf-8 on windows
-    # tqdm bails when trying to render the progress bar in this setup.
-    # because cp65001 is utf-8, we just map that codepage to the utf-8 codec.
-    # see #380 and: https://stackoverflow.com/a/3259271/87207
+    # Since Python 3.8 cp65001 is an alias to utf_8, but not for Pyhton < 3.8
+    # TODO: remove this code when only supporting Python 3.8+
+    # https://stackoverflow.com/a/3259271/87207
     import codecs
 
     codecs.register(lambda name: codecs.lookup("utf-8") if name == "cp65001" else None)
@@ -597,8 +757,52 @@ def handle_common_args(args):
     else:
         raise RuntimeError("unexpected --color value: " + args.color)
 
+    if hasattr(args, "rules"):
+        if args.rules == RULES_PATH_DEFAULT_STRING:
+            logger.debug("-" * 80)
+            logger.debug(" Using default embedded rules.")
+            logger.debug(" To provide your own rules, use the form `capa.exe -r ./path/to/rules/  /path/to/mal.exe`.")
+            logger.debug(" You can see the current default rule set here:")
+            logger.debug("     https://github.com/fireeye/capa-rules")
+            logger.debug("-" * 80)
+
+            rules_path = os.path.join(get_default_root(), "rules")
+
+            if not os.path.exists(rules_path):
+                # when a users installs capa via pip,
+                # this pulls down just the source code - not the default rules.
+                # i'm not sure the default rules should even be written to the library directory,
+                # so in this case, we require the user to use -r to specify the rule directory.
+                logger.error("default embedded rules not found! (maybe you installed capa as a library?)")
+                logger.error("provide your own rule set via the `-r` option.")
+                return -1
+        else:
+            rules_path = args.rules
+            logger.debug("using rules path: %s", rules_path)
+
+        args.rules = rules_path
+
+    if hasattr(args, "signatures"):
+        if args.signatures == SIGNATURES_PATH_DEFAULT_STRING:
+            logger.debug("-" * 80)
+            logger.debug(" Using default embedded signatures.")
+            logger.debug(
+                " To provide your own signatures, use the form `capa.exe --signature ./path/to/signatures/  /path/to/mal.exe`."
+            )
+            logger.debug("-" * 80)
+
+            sigs_path = os.path.join(get_default_root(), "sigs")
+        else:
+            sigs_path = args.signatures
+            logger.debug("using signatures path: %s", sigs_path)
+
+        args.signatures = sigs_path
+
 
 def main(argv=None):
+    if sys.version_info < (3, 6):
+        raise UnsupportedRuntimeError("This version of capa can only be used with Python 3.6+")
+
     if argv is None:
         argv = sys.argv[1:]
 
@@ -634,7 +838,7 @@ def main(argv=None):
     parser = argparse.ArgumentParser(
         description=desc, epilog=epilog, formatter_class=argparse.RawDescriptionHelpFormatter
     )
-    install_common_args(parser, {"sample", "format", "backend", "rules", "tag"})
+    install_common_args(parser, {"sample", "format", "backend", "signatures", "rules", "tag"})
     parser.add_argument("-j", "--json", action="store_true", help="emit JSON instead of text")
     args = parser.parse_args(args=argv)
     handle_common_args(args)
@@ -647,37 +851,8 @@ def main(argv=None):
         logger.error("%s", e.args[0])
         return -1
 
-    if args.rules == RULES_PATH_DEFAULT_STRING:
-        logger.debug("-" * 80)
-        logger.debug(" Using default embedded rules.")
-        logger.debug(" To provide your own rules, use the form `capa.exe -r ./path/to/rules/  /path/to/mal.exe`.")
-        logger.debug(" You can see the current default rule set here:")
-        logger.debug("     https://github.com/fireeye/capa-rules")
-        logger.debug("-" * 80)
-
-        if hasattr(sys, "frozen") and hasattr(sys, "_MEIPASS"):
-            logger.debug("detected running under PyInstaller")
-            rules_path = os.path.join(sys._MEIPASS, "rules")
-            logger.debug("default rule path (PyInstaller method): %s", rules_path)
-        else:
-            logger.debug("detected running from source")
-            rules_path = os.path.join(os.path.dirname(__file__), "..", "rules")
-            logger.debug("default rule path (source method): %s", rules_path)
-
-        if not os.path.exists(rules_path):
-            # when a users installs capa via pip,
-            # this pulls down just the source code - not the default rules.
-            # i'm not sure the default rules should even be written to the library directory,
-            # so in this case, we require the user to use -r to specify the rule directory.
-            logger.error("default embedded rules not found! (maybe you installed capa as a library?)")
-            logger.error("provide your own rule set via the `-r` option.")
-            return -1
-    else:
-        rules_path = args.rules
-        logger.debug("using rules path: %s", rules_path)
-
     try:
-        rules = get_rules(rules_path, disable_progress=args.quiet)
+        rules = get_rules(args.rules, disable_progress=args.quiet)
         rules = capa.rules.RuleSet(rules)
         logger.debug(
             "successfully loaded %s rules",
@@ -688,7 +863,7 @@ def main(argv=None):
         )
         if args.tag:
             rules = rules.filter_rules_by_meta(args.tag)
-            logger.debug("selected %s rules", len(rules))
+            logger.debug("selected %d rules", len(rules))
             for i, r in enumerate(rules.rules, 1):
                 # TODO don't display subscope rules?
                 logger.debug(" %d. %s", i, r)
@@ -696,15 +871,47 @@ def main(argv=None):
         logger.error("%s", str(e))
         return -1
 
+    if args.format == "pe" or (args.format == "auto" and taste.startswith(b"MZ")):
+        # this pefile file feature extractor is pretty light weight: it doesn't do any code analysis.
+        # so we can fairly quickly determine if the given PE file has "pure" file-scope rules
+        # that indicate a limitation (like "file is packed based on section names")
+        # and avoid doing a full code analysis on difficult/impossible binaries.
+        try:
+            from pefile import PEFormatError
+
+            file_extractor = capa.features.extractors.pefile.PefileFeatureExtractor(args.sample)
+        except PEFormatError as e:
+            logger.error("Input file '%s' is not a valid PE file: %s", args.sample, str(e))
+            return -1
+        pure_file_capabilities, _ = find_file_capabilities(rules, file_extractor, {})
+
+        # file limitations that rely on non-file scope won't be detected here.
+        # nor on FunctionName features, because pefile doesn't support this.
+        if has_file_limitation(rules, pure_file_capabilities):
+            # bail if capa encountered file limitation e.g. a packed binary
+            # do show the output in verbose mode, though.
+            if not (args.verbose or args.vverbose or args.json):
+                logger.debug("file limitation short circuit, won't analyze fully.")
+                return -1
+
+    try:
+        sig_paths = get_signatures(args.signatures)
+    except (IOError) as e:
+        logger.error("%s", str(e))
+        return -1
+
     if (args.format == "freeze") or (args.format == "auto" and capa.features.freeze.is_freeze(taste)):
         format = "freeze"
         with open(args.sample, "rb") as f:
             extractor = capa.features.freeze.load(f.read())
     else:
         format = args.format
+        should_save_workspace = os.environ.get("CAPA_SAVE_WORKSPACE") not in ("0", "no", "NO", "n", None)
+
         try:
-            backend = args.backend if sys.version_info > (3, 0) else BACKEND_VIV
-            extractor = get_extractor(args.sample, args.format, backend, disable_progress=args.quiet)
+            extractor = get_extractor(
+                args.sample, format, args.backend, sig_paths, should_save_workspace, disable_progress=args.quiet
+            )
         except UnsupportedFormatError:
             logger.error("-" * 80)
             logger.error(" Input file does not appear to be a PE file.")
@@ -715,16 +922,6 @@ def main(argv=None):
             logger.error(" If you don't know the input file type, you can try using the `file` utility to guess it.")
             logger.error("-" * 80)
             return -1
-        except UnsupportedRuntimeError:
-            logger.error("-" * 80)
-            logger.error(" Unsupported runtime or Python interpreter.")
-            logger.error(" ")
-            logger.error(" capa supports running under Python 2.7 using Vivisect for binary analysis.")
-            logger.error(" It can also run within IDA Pro, using either Python 2.7 or 3.5+.")
-            logger.error(" ")
-            logger.error(" If you're seeing this message on the command line, please ensure you're running Python 2.7.")
-            logger.error("-" * 80)
-            return -1
 
     meta = collect_metadata(argv, args.sample, args.rules, format, extractor)
 
@@ -738,13 +935,13 @@ def main(argv=None):
             return -1
 
     if args.json:
-        print(capa.render.render_json(meta, rules, capabilities))
+        print(capa.render.json.render(meta, rules, capabilities))
     elif args.vverbose:
-        print(capa.render.render_vverbose(meta, rules, capabilities))
+        print(capa.render.vverbose.render(meta, rules, capabilities))
     elif args.verbose:
-        print(capa.render.render_verbose(meta, rules, capabilities))
+        print(capa.render.verbose.render(meta, rules, capabilities))
     else:
-        print(capa.render.render_default(meta, rules, capabilities))
+        print(capa.render.default.render(meta, rules, capabilities))
     colorama.deinit()
 
     logger.debug("done.")
@@ -753,8 +950,10 @@ def main(argv=None):
 
 
 def ida_main():
+    import capa.rules
     import capa.ida.helpers
-    import capa.features.extractors.ida
+    import capa.render.default
+    import capa.features.extractors.ida.extractor
 
     logging.basicConfig(level=logging.INFO)
     logging.getLogger().setLevel(logging.INFO)
@@ -772,28 +971,21 @@ def ida_main():
     logger.debug("     https://github.com/fireeye/capa-rules")
     logger.debug("-" * 80)
 
-    if hasattr(sys, "frozen") and hasattr(sys, "_MEIPASS"):
-        logger.debug("detected running under PyInstaller")
-        rules_path = os.path.join(sys._MEIPASS, "rules")
-        logger.debug("default rule path (PyInstaller method): %s", rules_path)
-    else:
-        logger.debug("detected running from source")
-        rules_path = os.path.join(os.path.dirname(__file__), "..", "rules")
-        logger.debug("default rule path (source method): %s", rules_path)
-
+    rules_path = os.path.join(get_default_root(), "rules")
+    logger.debug("rule path: %s", rules_path)
     rules = get_rules(rules_path)
     rules = capa.rules.RuleSet(rules)
 
     meta = capa.ida.helpers.collect_metadata()
 
-    capabilities, counts = find_capabilities(rules, capa.features.extractors.ida.IdaFeatureExtractor())
+    capabilities, counts = find_capabilities(rules, capa.features.extractors.ida.extractor.IdaFeatureExtractor())
     meta["analysis"].update(counts)
 
     if has_file_limitation(rules, capabilities, is_standalone=False):
         capa.ida.helpers.inform_user_ida_ui("capa encountered warnings during analysis")
 
     colorama.init(strip=True)
-    print(capa.render.render_default(meta, rules, capabilities))
+    print(capa.render.default.render(meta, rules, capabilities))
 
 
 def is_runtime_ida():

capa/render/__init__.py

@@ -1,266 +0,0 @@
-# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
-# Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at: [package root]/LICENSE.txt
-# Unless required by applicable law or agreed to in writing, software distributed under the License
-#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and limitations under the License.
-
-import json
-
-import six
-
-import capa.rules
-import capa.engine
-
-
-def convert_statement_to_result_document(statement):
-    """
-    "statement": {
-        "type": "or"
-    },
-
-    "statement": {
-        "max": 9223372036854775808,
-        "min": 2,
-        "type": "range"
-    },
-    """
-    statement_type = statement.name.lower()
-    result = {"type": statement_type}
-    if statement.description:
-        result["description"] = statement.description
-
-    if statement_type == "some" and statement.count == 0:
-        result["type"] = "optional"
-    elif statement_type == "some":
-        result["count"] = statement.count
-    elif statement_type == "range":
-        result["min"] = statement.min
-        result["max"] = statement.max
-        result["child"] = convert_feature_to_result_document(statement.child)
-    elif statement_type == "subscope":
-        result["subscope"] = statement.scope
-
-    return result
-
-
-def convert_feature_to_result_document(feature):
-    """
-    "feature": {
-        "number": 6,
-        "type": "number"
-    },
-
-    "feature": {
-        "api": "ws2_32.WSASocket",
-        "type": "api"
-    },
-
-    "feature": {
-        "match": "create TCP socket",
-        "type": "match"
-    },
-
-    "feature": {
-        "characteristic": [
-            "loop",
-            true
-        ],
-        "type": "characteristic"
-    },
-    """
-    result = {"type": feature.name, feature.name: feature.get_value_str()}
-    if feature.description:
-        result["description"] = feature.description
-    if feature.name == "regex":
-        result["match"] = feature.match
-    return result
-
-
-def convert_node_to_result_document(node):
-    """
-    "node": {
-        "type": "statement",
-        "statement": { ... }
-    },
-
-    "node": {
-        "type": "feature",
-        "feature": { ... }
-    },
-    """
-
-    if isinstance(node, capa.engine.Statement):
-        return {
-            "type": "statement",
-            "statement": convert_statement_to_result_document(node),
-        }
-    elif isinstance(node, capa.features.Feature):
-        return {
-            "type": "feature",
-            "feature": convert_feature_to_result_document(node),
-        }
-    else:
-        raise RuntimeError("unexpected match node type")
-
-
-def convert_match_to_result_document(rules, capabilities, result):
-    """
-    convert the given Result instance into a common, Python-native data structure.
-    this will become part of the "result document" format that can be emitted to JSON.
-    """
-    doc = {
-        "success": bool(result.success),
-        "node": convert_node_to_result_document(result.statement),
-        "children": [convert_match_to_result_document(rules, capabilities, child) for child in result.children],
-    }
-
-    # logic expression, like `and`, don't have locations - their children do.
-    # so only add `locations` to feature nodes.
-    if isinstance(result.statement, capa.features.Feature):
-        if bool(result.success):
-            doc["locations"] = result.locations
-    elif isinstance(result.statement, capa.rules.Range):
-        if bool(result.success):
-            doc["locations"] = result.locations
-
-    # if we have a `match` statement, then we're referencing another rule.
-    # this could an external rule (written by a human), or
-    #  rule generated to support a subscope (basic block, etc.)
-    # we still want to include the matching logic in this tree.
-    #
-    # so, we need to lookup the other rule results
-    # and then filter those down to the address used here.
-    # finally, splice that logic into this tree.
-    if (
-        doc["node"]["type"] == "feature"
-        and doc["node"]["feature"]["type"] == "match"
-        # only add subtree on success,
-        # because there won't be results for the other rule on failure.
-        and doc["success"]
-    ):
-
-        rule_name = doc["node"]["feature"]["match"]
-        rule = rules[rule_name]
-        rule_matches = {address: result for (address, result) in capabilities[rule_name]}
-
-        if rule.meta.get("capa/subscope-rule"):
-            # for a subscope rule, fixup the node to be a scope node, rather than a match feature node.
-            #
-            # e.g. `contain loop/30c4c78e29bf4d54894fc74f664c62e8` -> `basic block`
-            scope = rule.meta["scope"]
-            doc["node"] = {
-                "type": "statement",
-                "statement": {
-                    "type": "subscope",
-                    "subscope": scope,
-                },
-            }
-
-        for location in doc["locations"]:
-            doc["children"].append(convert_match_to_result_document(rules, capabilities, rule_matches[location]))
-
-    return doc
-
-
-def convert_capabilities_to_result_document(meta, rules, capabilities):
-    """
-    convert the given rule set and capabilities result to a common, Python-native data structure.
-    this format can be directly emitted to JSON, or passed to the other `render_*` routines
-     to render as text.
-
-    see examples of substructures in above routines.
-
-    schema:
-
-    ```json
-    {
-      "meta": {...},
-      "rules: {
-        $rule-name: {
-          "meta": {...copied from rule.meta...},
-          "matches: {
-            $address: {...match details...},
-            ...
-          }
-        },
-        ...
-      }
-    }
-    ```
-
-    Args:
-      meta (Dict[str, Any]):
-      rules (RuleSet):
-      capabilities (Dict[str, List[Tuple[int, Result]]]):
-    """
-    doc = {
-        "meta": meta,
-        "rules": {},
-    }
-
-    for rule_name, matches in capabilities.items():
-        rule = rules[rule_name]
-
-        if rule.meta.get("capa/subscope-rule"):
-            continue
-
-        doc["rules"][rule_name] = {
-            "meta": dict(rule.meta),
-            "source": rule.definition,
-            "matches": {
-                addr: convert_match_to_result_document(rules, capabilities, match) for (addr, match) in matches
-            },
-        }
-
-    return doc
-
-
-def render_vverbose(meta, rules, capabilities):
-    # there's an import loop here
-    # if capa.render imports capa.render.vverbose
-    # and capa.render.vverbose import capa.render (implicitly, as a submodule)
-    # so, defer the import until routine is called, breaking the import loop.
-    import capa.render.vverbose
-
-    doc = convert_capabilities_to_result_document(meta, rules, capabilities)
-    return capa.render.vverbose.render_vverbose(doc)
-
-
-def render_verbose(meta, rules, capabilities):
-    # break import loop
-    import capa.render.verbose
-
-    doc = convert_capabilities_to_result_document(meta, rules, capabilities)
-    return capa.render.verbose.render_verbose(doc)
-
-
-def render_default(meta, rules, capabilities):
-    # break import loop
-    import capa.render.default
-    import capa.render.verbose
-
-    doc = convert_capabilities_to_result_document(meta, rules, capabilities)
-    return capa.render.default.render_default(doc)
-
-
-class CapaJsonObjectEncoder(json.JSONEncoder):
-    """JSON encoder that emits Python sets as sorted lists"""
-
-    def default(self, obj):
-        if isinstance(obj, (list, dict, int, float, bool, type(None))) or isinstance(obj, six.string_types):
-            return json.JSONEncoder.default(self, obj)
-        elif isinstance(obj, set):
-            return list(sorted(obj))
-        else:
-            # probably will TypeError
-            return json.JSONEncoder.default(self, obj)
-
-
-def render_json(meta, rules, capabilities):
-    return json.dumps(
-        convert_capabilities_to_result_document(meta, rules, capabilities),
-        cls=CapaJsonObjectEncoder,
-        sort_keys=True,
-    )

capa/render/default.py

@@ -7,16 +7,20 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 import collections
+from typing import Dict, List
 
-import six
 import tabulate
 
 import capa.render.utils as rutils
+import capa.render.result_document
+from capa.rules import RuleSet
+from capa.engine import MatchResults
+from capa.render.utils import StringIO
 
 tabulate.PRESERVE_WHITESPACE = True
 
 
-def width(s, character_count):
+def width(s: str, character_count: int) -> str:
     """pad the given string to at least `character_count`"""
     if len(s) < character_count:
         return s + " " * (character_count - len(s))
@@ -24,7 +28,7 @@ def width(s, character_count):
         return s
 
 
-def render_meta(doc, ostream):
+def render_meta(doc, ostream: StringIO):
     rows = [
         (width("md5", 22), width(doc["meta"]["sample"]["md5"], 82)),
         ("sha1", doc["meta"]["sample"]["sha1"]),
@@ -64,7 +68,7 @@ def find_subrule_matches(doc):
     return matches
 
 
-def render_capabilities(doc, ostream):
+def render_capabilities(doc, ostream: StringIO):
     """
     example::
 
@@ -102,7 +106,7 @@ def render_capabilities(doc, ostream):
         ostream.writeln(rutils.bold("no capabilities found"))
 
 
-def render_attack(doc, ostream):
+def render_attack(doc, ostream: StringIO):
     """
     example::
 
@@ -124,27 +128,16 @@ def render_attack(doc, ostream):
             continue
 
         for attack in rule["meta"]["att&ck"]:
-            tactic, _, rest = attack.partition("::")
-            if "::" in rest:
-                technique, _, rest = rest.partition("::")
-                subtechnique, _, id = rest.rpartition(" ")
-                tactics[tactic].add((technique, subtechnique, id))
-            else:
-                technique, _, id = rest.rpartition(" ")
-                tactics[tactic].add((technique, id))
+            tactics[attack["tactic"]].add((attack["technique"], attack.get("subtechnique"), attack["id"]))
 
     rows = []
     for tactic, techniques in sorted(tactics.items()):
         inner_rows = []
-        for spec in sorted(techniques):
-            if len(spec) == 2:
-                technique, id = spec
+        for (technique, subtechnique, id) in sorted(techniques):
+            if subtechnique is None:
                 inner_rows.append("%s %s" % (rutils.bold(technique), id))
-            elif len(spec) == 3:
-                technique, subtechnique, id = spec
-                inner_rows.append("%s::%s %s" % (rutils.bold(technique), subtechnique, id))
             else:
-                raise RuntimeError("unexpected ATT&CK spec format")
+                inner_rows.append("%s::%s %s" % (rutils.bold(technique), subtechnique, id))
         rows.append(
             (
                 rutils.bold(tactic.upper()),
@@ -161,7 +154,7 @@ def render_attack(doc, ostream):
         ostream.write("\n")
 
 
-def render_mbc(doc, ostream):
+def render_mbc(doc, ostream: StringIO):
     """
     example::
 
@@ -180,32 +173,17 @@ def render_mbc(doc, ostream):
         if not rule["meta"].get("mbc"):
             continue
 
-        mbcs = rule["meta"]["mbc"]
-        if not isinstance(mbcs, list):
-            raise ValueError("invalid rule: MBC mapping is not a list")
-
-        for mbc in mbcs:
-            objective, _, rest = mbc.partition("::")
-            if "::" in rest:
-                behavior, _, rest = rest.partition("::")
-                method, _, id = rest.rpartition(" ")
-                objectives[objective].add((behavior, method, id))
-            else:
-                behavior, _, id = rest.rpartition(" ")
-                objectives[objective].add((behavior, id))
+        for mbc in rule["meta"]["mbc"]:
+            objectives[mbc["objective"]].add((mbc["behavior"], mbc.get("method"), mbc["id"]))
 
     rows = []
     for objective, behaviors in sorted(objectives.items()):
         inner_rows = []
-        for spec in sorted(behaviors):
-            if len(spec) == 2:
-                behavior, id = spec
-                inner_rows.append("%s %s" % (rutils.bold(behavior), id))
-            elif len(spec) == 3:
-                behavior, method, id = spec
-                inner_rows.append("%s::%s %s" % (rutils.bold(behavior), method, id))
+        for (behavior, method, id) in sorted(behaviors):
+            if method is None:
+                inner_rows.append("%s [%s]" % (rutils.bold(behavior), id))
             else:
-                raise RuntimeError("unexpected MBC spec format")
+                inner_rows.append("%s::%s [%s]" % (rutils.bold(behavior), method, id))
         rows.append(
             (
                 rutils.bold(objective.upper()),
@@ -232,3 +210,8 @@ def render_default(doc):
     render_capabilities(doc, ostream)
 
     return ostream.getvalue()
+
+
+def render(meta, rules: RuleSet, capabilities: MatchResults) -> str:
+    doc = capa.render.result_document.convert_capabilities_to_result_document(meta, rules, capabilities)
+    return render_default(doc)

capa/render/json.py

@@ -0,0 +1,33 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+import json
+
+import capa.render.result_document
+from capa.rules import RuleSet
+from capa.engine import MatchResults
+
+
+class CapaJsonObjectEncoder(json.JSONEncoder):
+    """JSON encoder that emits Python sets as sorted lists"""
+
+    def default(self, obj):
+        if isinstance(obj, (list, dict, int, float, bool, type(None))) or isinstance(obj, str):
+            return json.JSONEncoder.default(self, obj)
+        elif isinstance(obj, set):
+            return list(sorted(obj))
+        else:
+            # probably will TypeError
+            return json.JSONEncoder.default(self, obj)
+
+
+def render(meta, rules: RuleSet, capabilities: MatchResults) -> str:
+    return json.dumps(
+        capa.render.result_document.convert_capabilities_to_result_document(meta, rules, capabilities),
+        cls=CapaJsonObjectEncoder,
+        sort_keys=True,
+    )

capa/render/result_document.py

@@ -0,0 +1,319 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+import copy
+
+import capa.rules
+import capa.engine
+import capa.render.utils
+import capa.features.common
+from capa.rules import RuleSet
+from capa.engine import MatchResults
+
+
+def convert_statement_to_result_document(statement):
+    """
+    "statement": {
+        "type": "or"
+    },
+
+    "statement": {
+        "max": 9223372036854775808,
+        "min": 2,
+        "type": "range"
+    },
+    """
+    statement_type = statement.name.lower()
+    result = {"type": statement_type}
+    if statement.description:
+        result["description"] = statement.description
+
+    if statement_type == "some" and statement.count == 0:
+        result["type"] = "optional"
+    elif statement_type == "some":
+        result["count"] = statement.count
+    elif statement_type == "range":
+        result["min"] = statement.min
+        result["max"] = statement.max
+        result["child"] = convert_feature_to_result_document(statement.child)
+    elif statement_type == "subscope":
+        result["subscope"] = statement.scope
+
+    return result
+
+
+def convert_feature_to_result_document(feature):
+    """
+    "feature": {
+        "number": 6,
+        "type": "number"
+    },
+
+    "feature": {
+        "api": "ws2_32.WSASocket",
+        "type": "api"
+    },
+
+    "feature": {
+        "match": "create TCP socket",
+        "type": "match"
+    },
+
+    "feature": {
+        "characteristic": [
+            "loop",
+            true
+        ],
+        "type": "characteristic"
+    },
+    """
+    result = {"type": feature.name, feature.name: feature.get_value_str()}
+    if feature.description:
+        result["description"] = feature.description
+    if feature.name == "regex":
+        result["matches"] = feature.matches
+    return result
+
+
+def convert_node_to_result_document(node):
+    """
+    "node": {
+        "type": "statement",
+        "statement": { ... }
+    },
+
+    "node": {
+        "type": "feature",
+        "feature": { ... }
+    },
+    """
+
+    if isinstance(node, capa.engine.Statement):
+        return {
+            "type": "statement",
+            "statement": convert_statement_to_result_document(node),
+        }
+    elif isinstance(node, capa.features.common.Feature):
+        return {
+            "type": "feature",
+            "feature": convert_feature_to_result_document(node),
+        }
+    else:
+        raise RuntimeError("unexpected match node type")
+
+
+def convert_match_to_result_document(rules, capabilities, result):
+    """
+    convert the given Result instance into a common, Python-native data structure.
+    this will become part of the "result document" format that can be emitted to JSON.
+    """
+    doc = {
+        "success": bool(result.success),
+        "node": convert_node_to_result_document(result.statement),
+        "children": [convert_match_to_result_document(rules, capabilities, child) for child in result.children],
+    }
+
+    # logic expression, like `and`, don't have locations - their children do.
+    # so only add `locations` to feature nodes.
+    if isinstance(result.statement, capa.features.common.Feature):
+        if bool(result.success):
+            doc["locations"] = result.locations
+    elif isinstance(result.statement, capa.engine.Range):
+        if bool(result.success):
+            doc["locations"] = result.locations
+
+    # if we have a `match` statement, then we're referencing another rule or namespace.
+    # this could an external rule (written by a human), or
+    #  rule generated to support a subscope (basic block, etc.)
+    # we still want to include the matching logic in this tree.
+    #
+    # so, we need to lookup the other rule results
+    # and then filter those down to the address used here.
+    # finally, splice that logic into this tree.
+    if (
+        doc["node"]["type"] == "feature"
+        and doc["node"]["feature"]["type"] == "match"
+        # only add subtree on success,
+        # because there won't be results for the other rule on failure.
+        and doc["success"]
+    ):
+
+        name = doc["node"]["feature"]["match"]
+
+        if name in rules:
+            # this is a rule that we're matching
+            #
+            # pull matches from the referenced rule into our tree here.
+            rule_name = doc["node"]["feature"]["match"]
+            rule = rules[rule_name]
+            rule_matches = {address: result for (address, result) in capabilities[rule_name]}
+
+            if rule.meta.get("capa/subscope-rule"):
+                # for a subscope rule, fixup the node to be a scope node, rather than a match feature node.
+                #
+                # e.g. `contain loop/30c4c78e29bf4d54894fc74f664c62e8` -> `basic block`
+                scope = rule.meta["scope"]
+                doc["node"] = {
+                    "type": "statement",
+                    "statement": {
+                        "type": "subscope",
+                        "subscope": scope,
+                    },
+                }
+
+            for location in doc["locations"]:
+                doc["children"].append(convert_match_to_result_document(rules, capabilities, rule_matches[location]))
+        else:
+            # this is a namespace that we're matching
+            #
+            # check for all rules in the namespace,
+            # seeing if they matched.
+            # if so, pull their matches into our match tree here.
+            ns_name = doc["node"]["feature"]["match"]
+            ns_rules = rules.rules_by_namespace[ns_name]
+
+            for rule in ns_rules:
+                if rule.name in capabilities:
+                    # the rule matched, so splice results into our tree here.
+                    #
+                    # note, there's a shortcoming in our result document schema here:
+                    # we lose the name of the rule that matched in a namespace.
+                    # for example, if we have a statement: `match: runtime/dotnet`
+                    # and we get matches, we can say the following:
+                    #
+                    #     match: runtime/dotnet @ 0x0
+                    #       or:
+                    #         import: mscoree._CorExeMain @ 0x402000
+                    #
+                    # however, we lose the fact that it was rule
+                    #   "compiled to the .NET platform"
+                    # that contained this logic and did the match.
+                    #
+                    # we could introduce an intermediate node here.
+                    # this would be a breaking change and require updates to the renderers.
+                    # in the meantime, the above might be sufficient.
+                    rule_matches = {address: result for (address, result) in capabilities[rule.name]}
+                    for location in doc["locations"]:
+                        doc["children"].append(
+                            convert_match_to_result_document(rules, capabilities, rule_matches[location])
+                        )
+
+    return doc
+
+
+def convert_meta_to_result_document(meta):
+    # make a copy so that we don't modify the given parameter
+    meta = copy.deepcopy(meta)
+
+    attacks = meta.get("att&ck", [])
+    meta["att&ck"] = [parse_canonical_attack(attack) for attack in attacks]
+    mbcs = meta.get("mbc", [])
+    meta["mbc"] = [parse_canonical_mbc(mbc) for mbc in mbcs]
+    return meta
+
+
+def parse_canonical_attack(attack: str):
+    """
+    parse capa's canonical ATT&CK representation: `Tactic::Technique::Subtechnique [Identifier]`
+    """
+    tactic = ""
+    technique = ""
+    subtechnique = ""
+    parts, id = capa.render.utils.parse_parts_id(attack)
+    if len(parts) > 0:
+        tactic = parts[0]
+    if len(parts) > 1:
+        technique = parts[1]
+    if len(parts) > 2:
+        subtechnique = parts[2]
+
+    return {
+        "parts": parts,
+        "id": id,
+        "tactic": tactic,
+        "technique": technique,
+        "subtechnique": subtechnique,
+    }
+
+
+def parse_canonical_mbc(mbc: str):
+    """
+    parse capa's canonical MBC representation: `Objective::Behavior::Method [Identifier]`
+    """
+    objective = ""
+    behavior = ""
+    method = ""
+    parts, id = capa.render.utils.parse_parts_id(mbc)
+    if len(parts) > 0:
+        objective = parts[0]
+    if len(parts) > 1:
+        behavior = parts[1]
+    if len(parts) > 2:
+        method = parts[2]
+
+    return {
+        "parts": parts,
+        "id": id,
+        "objective": objective,
+        "behavior": behavior,
+        "method": method,
+    }
+
+
+def convert_capabilities_to_result_document(meta, rules: RuleSet, capabilities: MatchResults):
+    """
+    convert the given rule set and capabilities result to a common, Python-native data structure.
+    this format can be directly emitted to JSON, or passed to the other `capa.render.*.render()` routines
+     to render as text.
+
+    see examples of substructures in above routines.
+
+    schema:
+
+    ```json
+    {
+      "meta": {...},
+      "rules: {
+        $rule-name: {
+          "meta": {...copied from rule.meta...},
+          "matches: {
+            $address: {...match details...},
+            ...
+          }
+        },
+        ...
+      }
+    }
+    ```
+
+    Args:
+      meta (Dict[str, Any]):
+      rules (RuleSet):
+      capabilities (Dict[str, List[Tuple[int, Result]]]):
+    """
+    doc = {
+        "meta": meta,
+        "rules": {},
+    }
+
+    for rule_name, matches in capabilities.items():
+        rule = rules[rule_name]
+
+        if rule.meta.get("capa/subscope-rule"):
+            continue
+
+        rule_meta = convert_meta_to_result_document(rule.meta)
+
+        doc["rules"][rule_name] = {
+            "meta": rule_meta,
+            "source": rule.definition,
+            "matches": {
+                addr: convert_match_to_result_document(rules, capabilities, match) for (addr, match) in matches
+            },
+        }
+
+    return doc

capa/render/utils.py

@@ -6,21 +6,22 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import six
+import io
+
 import termcolor
 
 
-def bold(s):
+def bold(s: str) -> str:
     """draw attention to the given string"""
     return termcolor.colored(s, "blue")
 
 
-def bold2(s):
+def bold2(s: str) -> str:
     """draw attention to the given string, within a `bold` section"""
     return termcolor.colored(s, "green")
 
 
-def hex(n):
+def hex(n: int) -> str:
     """render the given number using upper case hex, like: 0x123ABC"""
     if n < 0:
         return "-0x%X" % (-n)
@@ -28,6 +29,24 @@ def hex(n):
         return "0x%X" % n
 
 
+def parse_parts_id(s: str):
+    id = ""
+    parts = s.split("::")
+    if len(parts) > 0:
+        last = parts.pop()
+        last, _, id = last.rpartition(" ")
+        id = id.lstrip("[").rstrip("]")
+        parts.append(last)
+    return parts, id
+
+
+def format_parts_id(data):
+    """
+    format canonical representation of ATT&CK/MBC parts and ID
+    """
+    return "%s [%s]" % ("::".join(data["parts"]), data["id"])
+
+
 def capability_rules(doc):
     """enumerate the rules in (namespace, name) order that are 'capability' rules (not lib/subscope/disposition/etc)."""
     for (_, _, rule) in sorted(
@@ -49,7 +68,7 @@ def capability_rules(doc):
         yield rule
 
 
-class StringIO(six.StringIO):
+class StringIO(io.StringIO):
     def writeln(self, s):
         self.write(s)
         self.write("\n")

capa/render/verbose.py

@@ -26,6 +26,9 @@ import tabulate
 
 import capa.rules
 import capa.render.utils as rutils
+import capa.render.result_document
+from capa.rules import RuleSet
+from capa.engine import MatchResults
 
 
 def render_meta(ostream, doc):
@@ -57,6 +60,7 @@ def render_meta(ostream, doc):
         ("base address", hex(doc["meta"]["analysis"]["base_address"])),
         ("rules", doc["meta"]["analysis"]["rules"]),
         ("function count", len(doc["meta"]["analysis"]["feature_counts"]["functions"])),
+        ("library function count", len(doc["meta"]["analysis"]["library_functions"])),
         (
             "total feature count",
             doc["meta"]["analysis"]["feature_counts"]["file"]
@@ -119,3 +123,8 @@ def render_verbose(doc):
     ostream.write("\n")
 
     return ostream.getvalue()
+
+
+def render(meta, rules: RuleSet, capabilities: MatchResults) -> str:
+    doc = capa.render.result_document.convert_capabilities_to_result_document(meta, rules, capabilities)
+    return render_verbose(doc)

capa/render/vverbose.py

@@ -6,13 +6,15 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import collections
-
 import tabulate
 
 import capa.rules
 import capa.render.utils as rutils
 import capa.render.verbose
+import capa.features.common
+import capa.render.result_document
+from capa.rules import RuleSet
+from capa.engine import MatchResults
 
 
 def render_locations(ostream, match):
@@ -56,7 +58,11 @@ def render_statement(ostream, match, statement, indent=0):
         child = statement["child"]
 
         if child[child["type"]]:
-            value = rutils.bold2(child[child["type"]])
+            if child["type"] == "string":
+                value = '"%s"' % capa.features.common.escape_string(child[child["type"]])
+            else:
+                value = child[child["type"]]
+            value = rutils.bold2(value)
             if child.get("description"):
                 ostream.write("count(%s(%s = %s)): " % (child["type"], value, child["description"]))
             else:
@@ -81,27 +87,50 @@ def render_statement(ostream, match, statement, indent=0):
         raise RuntimeError("unexpected match statement type: " + str(statement))
 
 
+def render_string_value(s):
+    return '"%s"' % capa.features.common.escape_string(s)
+
+
 def render_feature(ostream, match, feature, indent=0):
     ostream.write("  " * indent)
 
     key = feature["type"]
     value = feature[feature["type"]]
-    if key == "regex":
-        key = "string"  # render string for regex to mirror the rule source
-        value = feature["match"]  # the match provides more information than the value for regex
 
-    ostream.write(key)
-    ostream.write(": ")
+    if key != "regex":
+        # like:
+        #   number: 10 = SOME_CONSTANT @ 0x401000
+        if key == "string":
+            value = render_string_value(value)
 
-    if value:
-        ostream.write(rutils.bold2(value))
+        ostream.write(key)
+        ostream.write(": ")
 
-        if "description" in feature:
-            ostream.write(capa.rules.DESCRIPTION_SEPARATOR)
-            ostream.write(feature["description"])
+        if value:
+            ostream.write(rutils.bold2(value))
 
-    render_locations(ostream, match)
-    ostream.write("\n")
+            if "description" in feature:
+                ostream.write(capa.rules.DESCRIPTION_SEPARATOR)
+                ostream.write(feature["description"])
+
+        render_locations(ostream, match)
+        ostream.write("\n")
+    else:
+        # like:
+        #  regex: /blah/ = SOME_CONSTANT
+        #    - "foo blah baz" @ 0x401000
+        #    - "aaa blah bbb" @ 0x402000, 0x403400
+        ostream.write(key)
+        ostream.write(": ")
+        ostream.write(value)
+        ostream.write("\n")
+
+        for match, locations in sorted(feature["matches"].items(), key=lambda p: p[0]):
+            ostream.write("  " * (indent + 1))
+            ostream.write("- ")
+            ostream.write(rutils.bold2(render_string_value(match)))
+            render_locations(ostream, {"locations": locations})
+            ostream.write("\n")
 
 
 def render_node(ostream, match, node, indent=0):
@@ -190,6 +219,12 @@ def render_rules(ostream, doc):
                 continue
 
             v = rule["meta"][key]
+            if not v:
+                continue
+
+            if key in ("att&ck", "mbc"):
+                v = [rutils.format_parts_id(vv) for vv in v]
+
             if isinstance(v, list) and len(v) == 1:
                 v = v[0]
             elif isinstance(v, list) and len(v) > 1:
@@ -229,3 +264,8 @@ def render_vverbose(doc):
     ostream.write("\n")
 
     return ostream.getvalue()
+
+
+def render(meta, rules: RuleSet, capabilities: MatchResults) -> str:
+    doc = capa.render.result_document.convert_capabilities_to_result_document(meta, rules, capabilities)
+    return render_vverbose(doc)

capa/rules.py

@@ -6,29 +6,35 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
+import io
 import re
 import uuid
 import codecs
 import logging
 import binascii
 import functools
+import collections
 
 try:
     from functools import lru_cache
 except ImportError:
-    from backports.functools_lru_cache import lru_cache
+    # need to type ignore this due to mypy bug here (duplicate name):
+    # https://github.com/python/mypy/issues/1153
+    from backports.functools_lru_cache import lru_cache  # type: ignore
+
+from typing import Any, Set, Dict, List, Union, Iterator
 
-import six
 import yaml
 import ruamel.yaml
 
-import capa.engine
+import capa.engine as ceng
 import capa.features
 import capa.features.file
 import capa.features.insn
+import capa.features.common
 import capa.features.basicblock
-from capa.engine import *
-from capa.features import MAX_BYTES_FEATURE_SIZE
+from capa.engine import Statement, FeatureSet
+from capa.features.common import MAX_BYTES_FEATURE_SIZE, Feature
 
 logger = logging.getLogger(__name__)
 
@@ -65,37 +71,38 @@ BASIC_BLOCK_SCOPE = "basic block"
 
 SUPPORTED_FEATURES = {
     FILE_SCOPE: {
-        capa.features.MatchedRule,
+        capa.features.common.MatchedRule,
         capa.features.file.Export,
         capa.features.file.Import,
         capa.features.file.Section,
-        capa.features.Characteristic("embedded pe"),
-        capa.features.String,
+        capa.features.file.FunctionName,
+        capa.features.common.Characteristic("embedded pe"),
+        capa.features.common.String,
     },
     FUNCTION_SCOPE: {
         # plus basic block scope features, see below
         capa.features.basicblock.BasicBlock,
-        capa.features.Characteristic("calls from"),
-        capa.features.Characteristic("calls to"),
-        capa.features.Characteristic("loop"),
-        capa.features.Characteristic("recursive call"),
+        capa.features.common.Characteristic("calls from"),
+        capa.features.common.Characteristic("calls to"),
+        capa.features.common.Characteristic("loop"),
+        capa.features.common.Characteristic("recursive call"),
     },
     BASIC_BLOCK_SCOPE: {
-        capa.features.MatchedRule,
+        capa.features.common.MatchedRule,
         capa.features.insn.API,
         capa.features.insn.Number,
-        capa.features.String,
-        capa.features.Bytes,
+        capa.features.common.String,
+        capa.features.common.Bytes,
         capa.features.insn.Offset,
         capa.features.insn.Mnemonic,
-        capa.features.Characteristic("nzxor"),
-        capa.features.Characteristic("peb access"),
-        capa.features.Characteristic("fs access"),
-        capa.features.Characteristic("gs access"),
-        capa.features.Characteristic("cross section flow"),
-        capa.features.Characteristic("tight loop"),
-        capa.features.Characteristic("stack string"),
-        capa.features.Characteristic("indirect call"),
+        capa.features.common.Characteristic("nzxor"),
+        capa.features.common.Characteristic("peb access"),
+        capa.features.common.Characteristic("fs access"),
+        capa.features.common.Characteristic("gs access"),
+        capa.features.common.Characteristic("cross section flow"),
+        capa.features.common.Characteristic("tight loop"),
+        capa.features.common.Characteristic("stack string"),
+        capa.features.common.Characteristic("indirect call"),
     },
 }
 
@@ -138,22 +145,32 @@ class InvalidRuleSet(ValueError):
         return str(self)
 
 
-def ensure_feature_valid_for_scope(scope, feature):
-    if isinstance(feature, capa.features.Characteristic):
-        if capa.features.Characteristic(feature.value) not in SUPPORTED_FEATURES[scope]:
-            raise InvalidRule("feature %s not support for scope %s" % (feature, scope))
-    elif not isinstance(feature, tuple(filter(lambda t: isinstance(t, type), SUPPORTED_FEATURES[scope]))):
+def ensure_feature_valid_for_scope(scope: str, feature: Union[Feature, Statement]):
+    # if the given feature is a characteristic,
+    # check that is a valid characteristic for the given scope.
+    if (
+        isinstance(feature, capa.features.common.Characteristic)
+        and isinstance(feature.value, str)
+        and capa.features.common.Characteristic(feature.value) not in SUPPORTED_FEATURES[scope]
+    ):
         raise InvalidRule("feature %s not support for scope %s" % (feature, scope))
 
+    if not isinstance(feature, capa.features.common.Characteristic):
+        # features of this scope that are not Characteristics will be Type instances.
+        # check that the given feature is one of these types.
+        types_for_scope = filter(lambda t: isinstance(t, type), SUPPORTED_FEATURES[scope])
+        if not isinstance(feature, tuple(types_for_scope)):  # type: ignore
+            raise InvalidRule("feature %s not support for scope %s" % (feature, scope))
 
-def parse_int(s):
+
+def parse_int(s: str) -> int:
     if s.startswith("0x"):
         return int(s, 0x10)
     else:
         return int(s, 10)
 
 
-def parse_range(s):
+def parse_range(s: str):
     """
     parse a string "(0, 1)" into a range (min, max).
     min and/or max may by None to indicate an unbound range.
@@ -166,23 +183,21 @@ def parse_range(s):
         raise InvalidRule("invalid range: %s" % (s))
 
     s = s[len("(") : -len(")")]
-    min, _, max = s.partition(",")
-    min = min.strip()
-    max = max.strip()
+    min_spec, _, max_spec = s.partition(",")
+    min_spec = min_spec.strip()
+    max_spec = max_spec.strip()
 
-    if min:
-        min = parse_int(min.strip())
+    min = None
+    if min_spec:
+        min = parse_int(min_spec)
         if min < 0:
             raise InvalidRule("range min less than zero")
-    else:
-        min = None
 
-    if max:
-        max = parse_int(max.strip())
+    max = None
+    if max_spec:
+        max = parse_int(max_spec)
         if max < 0:
             raise InvalidRule("range max less than zero")
-    else:
-        max = None
 
     if min is not None and max is not None:
         if max < min:
@@ -191,14 +206,14 @@ def parse_range(s):
     return min, max
 
 
-def parse_feature(key):
+def parse_feature(key: str):
     # keep this in sync with supported features
     if key == "api":
         return capa.features.insn.API
     elif key == "string":
-        return capa.features.StringFactory
+        return capa.features.common.StringFactory
     elif key == "bytes":
-        return capa.features.Bytes
+        return capa.features.common.Bytes
     elif key == "number":
         return capa.features.insn.Number
     elif key.startswith("number/"):
@@ -220,7 +235,7 @@ def parse_feature(key):
     elif key == "basic blocks":
         return capa.features.basicblock.BasicBlock
     elif key == "characteristic":
-        return capa.features.Characteristic
+        return capa.features.common.Characteristic
     elif key == "export":
         return capa.features.file.Export
     elif key == "import":
@@ -228,7 +243,9 @@ def parse_feature(key):
     elif key == "section":
         return capa.features.file.Section
     elif key == "match":
-        return capa.features.MatchedRule
+        return capa.features.common.MatchedRule
+    elif key == "function-name":
+        return capa.features.file.FunctionName
     else:
         raise InvalidRule("unexpected statement: %s" % key)
 
@@ -240,39 +257,70 @@ def parse_feature(key):
 DESCRIPTION_SEPARATOR = " = "
 
 
-def parse_description(s, value_type, description=None):
-    """
-    s can be an int or a string
-    """
-    if value_type != "string" and isinstance(s, six.string_types) and DESCRIPTION_SEPARATOR in s:
-        if description:
-            raise InvalidRule(
-                'unexpected value: "%s", only one description allowed (inline description with `%s`)'
-                % (s, DESCRIPTION_SEPARATOR)
-            )
-        value, _, description = s.partition(DESCRIPTION_SEPARATOR)
-        if description == "":
-            raise InvalidRule('unexpected value: "%s", description cannot be empty' % s)
-    else:
+def parse_bytes(s: str) -> bytes:
+    try:
+        b = codecs.decode(s.replace(" ", "").encode("ascii"), "hex")
+    except binascii.Error:
+        raise InvalidRule('unexpected bytes value: must be a valid hex sequence: "%s"' % s)
+
+    if len(b) > MAX_BYTES_FEATURE_SIZE:
+        raise InvalidRule(
+            "unexpected bytes value: byte sequences must be no larger than %s bytes" % MAX_BYTES_FEATURE_SIZE
+        )
+
+    return b
+
+
+def parse_description(s: Union[str, int, bytes], value_type: str, description=None):
+    if value_type == "string":
+        # string features cannot have inline descriptions,
+        # so we assume the entire value is the string,
+        # like: `string: foo = bar` -> "foo = bar"
         value = s
+    else:
+        # other features can have inline descriptions, like `number: 10 = CONST_FOO`.
+        # in this case, the RHS will be like `10 = CONST_FOO` or some other string
+        if isinstance(s, str):
+            if DESCRIPTION_SEPARATOR in s:
+                if description:
+                    # there is already a description passed in as a sub node, like:
+                    #
+                    #     - number: 10 = CONST_FOO
+                    #       description: CONST_FOO
+                    raise InvalidRule(
+                        'unexpected value: "%s", only one description allowed (inline description with `%s`)'
+                        % (s, DESCRIPTION_SEPARATOR)
+                    )
 
-    if isinstance(value, six.string_types):
-        if value_type == "bytes":
-            try:
-                value = codecs.decode(value.replace(" ", ""), "hex")
-            # TODO: Remove TypeError when Python2 is not used anymore
-            except (TypeError, binascii.Error):
-                raise InvalidRule('unexpected bytes value: "%s", must be a valid hex sequence' % value)
+                value, _, description = s.partition(DESCRIPTION_SEPARATOR)
+                if description == "":
+                    # sanity check:
+                    # there is an empty description, like `number: 10 =`
+                    raise InvalidRule('unexpected value: "%s", description cannot be empty' % s)
+            else:
+                # this is a string, but there is no description,
+                # like: `api: CreateFileA`
+                value = s
 
-            if len(value) > MAX_BYTES_FEATURE_SIZE:
-                raise InvalidRule(
-                    "unexpected bytes value: byte sequences must be no larger than %s bytes" % MAX_BYTES_FEATURE_SIZE
-                )
-        elif value_type in ("number", "offset") or value_type.startswith(("number/", "offset/")):
-            try:
-                value = parse_int(value)
-            except ValueError:
-                raise InvalidRule('unexpected value: "%s", must begin with numerical value' % value)
+            # cast from the received string value to the appropriate type.
+            #
+            # without a description, this type would already be correct,
+            # but since we parsed the description from a string,
+            # we need to convert the value to the expected type.
+            #
+            # for example, from `number: 10 = CONST_FOO` we have
+            # the string "10" that needs to become the number 10.
+            if value_type == "bytes":
+                value = parse_bytes(value)
+            elif value_type in ("number", "offset") or value_type.startswith(("number/", "offset/")):
+                try:
+                    value = parse_int(value)
+                except ValueError:
+                    raise InvalidRule('unexpected value: "%s", must begin with numerical value' % value)
+
+        else:
+            # the value might be a number, like: `number: 10`
+            value = s
 
     return value, description
 
@@ -312,28 +360,28 @@ def pop_statement_description_entry(d):
     return description["description"]
 
 
-def build_statements(d, scope):
+def build_statements(d, scope: str):
     if len(d.keys()) > 2:
         raise InvalidRule("too many statements")
 
     key = list(d.keys())[0]
     description = pop_statement_description_entry(d[key])
     if key == "and":
-        return And([build_statements(dd, scope) for dd in d[key]], description=description)
+        return ceng.And([build_statements(dd, scope) for dd in d[key]], description=description)
     elif key == "or":
-        return Or([build_statements(dd, scope) for dd in d[key]], description=description)
+        return ceng.Or([build_statements(dd, scope) for dd in d[key]], description=description)
     elif key == "not":
         if len(d[key]) != 1:
             raise InvalidRule("not statement must have exactly one child statement")
-        return Not(build_statements(d[key][0], scope), description=description)
+        return ceng.Not(build_statements(d[key][0], scope), description=description)
     elif key.endswith(" or more"):
         count = int(key[: -len("or more")])
-        return Some(count, [build_statements(dd, scope) for dd in d[key]], description=description)
+        return ceng.Some(count, [build_statements(dd, scope) for dd in d[key]], description=description)
     elif key == "optional":
         # `optional` is an alias for `0 or more`
         # which is useful for documenting behaviors,
         # like with `write file`, we might say that `WriteFile` is optionally found alongside `CreateFileA`.
-        return Some(0, [build_statements(dd, scope) for dd in d[key]], description=description)
+        return ceng.Some(0, [build_statements(dd, scope) for dd in d[key]], description=description)
 
     elif key == "function":
         if scope != FILE_SCOPE:
@@ -342,7 +390,7 @@ def build_statements(d, scope):
         if len(d[key]) != 1:
             raise InvalidRule("subscope must have exactly one child statement")
 
-        return Subscope(FUNCTION_SCOPE, build_statements(d[key][0], FUNCTION_SCOPE))
+        return ceng.Subscope(FUNCTION_SCOPE, build_statements(d[key][0], FUNCTION_SCOPE))
 
     elif key == "basic block":
         if scope != FUNCTION_SCOPE:
@@ -351,7 +399,7 @@ def build_statements(d, scope):
         if len(d[key]) != 1:
             raise InvalidRule("subscope must have exactly one child statement")
 
-        return Subscope(BASIC_BLOCK_SCOPE, build_statements(d[key][0], BASIC_BLOCK_SCOPE))
+        return ceng.Subscope(BASIC_BLOCK_SCOPE, build_statements(d[key][0], BASIC_BLOCK_SCOPE))
 
     elif key.startswith("count(") and key.endswith(")"):
         # e.g.:
@@ -392,21 +440,21 @@ def build_statements(d, scope):
 
         count = d[key]
         if isinstance(count, int):
-            return Range(feature, min=count, max=count, description=description)
+            return ceng.Range(feature, min=count, max=count, description=description)
         elif count.endswith(" or more"):
             min = parse_int(count[: -len(" or more")])
             max = None
-            return Range(feature, min=min, max=max, description=description)
+            return ceng.Range(feature, min=min, max=max, description=description)
         elif count.endswith(" or fewer"):
             min = None
             max = parse_int(count[: -len(" or fewer")])
-            return Range(feature, min=min, max=max, description=description)
+            return ceng.Range(feature, min=min, max=max, description=description)
         elif count.startswith("("):
             min, max = parse_range(count)
-            return Range(feature, min=min, max=max, description=description)
+            return ceng.Range(feature, min=min, max=max, description=description)
         else:
             raise InvalidRule("unexpected range: %s" % (count))
-    elif key == "string" and not isinstance(d[key], six.string_types):
+    elif key == "string" and not isinstance(d[key], str):
         raise InvalidRule("ambiguous string value %s, must be defined as explicit string" % d[key])
     else:
         Feature = parse_feature(key)
@@ -419,16 +467,16 @@ def build_statements(d, scope):
         return feature
 
 
-def first(s):
+def first(s: List[Any]) -> Any:
     return s[0]
 
 
-def second(s):
+def second(s: List[Any]) -> Any:
     return s[1]
 
 
-class Rule(object):
-    def __init__(self, name, scope, statement, meta, definition=""):
+class Rule:
+    def __init__(self, name: str, scope: str, statement: Statement, meta, definition=""):
         super(Rule, self).__init__()
         self.name = name
         self.scope = scope
@@ -458,7 +506,7 @@ class Rule(object):
         deps = set([])
 
         def rec(statement):
-            if isinstance(statement, capa.features.MatchedRule):
+            if isinstance(statement, capa.features.common.MatchedRule):
                 # we're not sure at this point if the `statement.value` is
                 #  really a rule name or a namespace name (we use `MatchedRule` for both cases).
                 # we'll give precedence to namespaces, and then assume if that does work,
@@ -474,7 +522,7 @@ class Rule(object):
                     # not a namespace, assume its a rule name.
                     deps.add(statement.value)
 
-            elif isinstance(statement, Statement):
+            elif isinstance(statement, ceng.Statement):
                 for child in statement.get_children():
                     rec(child)
 
@@ -485,11 +533,9 @@ class Rule(object):
         return deps
 
     def _extract_subscope_rules_rec(self, statement):
-        if isinstance(statement, Statement):
+        if isinstance(statement, ceng.Statement):
             # for each child that is a subscope,
-            for subscope in filter(
-                lambda statement: isinstance(statement, capa.engine.Subscope), statement.get_children()
-            ):
+            for subscope in filter(lambda statement: isinstance(statement, ceng.Subscope), statement.get_children()):
 
                 # create a new rule from it.
                 # the name is a randomly generated, hopefully unique value.
@@ -514,7 +560,7 @@ class Rule(object):
                 )
 
                 # update the existing statement to `match` the new rule
-                new_node = capa.features.MatchedRule(name)
+                new_node = capa.features.common.MatchedRule(name)
                 statement.replace_child(subscope, new_node)
 
                 # and yield the new rule to our caller
@@ -551,15 +597,16 @@ class Rule(object):
         for new_rule in self._extract_subscope_rules_rec(self.statement):
             yield new_rule
 
-    def evaluate(self, features):
+    def evaluate(self, features: FeatureSet):
         return self.statement.evaluate(features)
 
     @classmethod
     def from_dict(cls, d, definition):
-        name = d["rule"]["meta"]["name"]
+        meta = d["rule"]["meta"]
+        name = meta["name"]
         # if scope is not specified, default to function scope.
         # this is probably the mode that rule authors will start with.
-        scope = d["rule"]["meta"].get("scope", FUNCTION_SCOPE)
+        scope = meta.get("scope", FUNCTION_SCOPE)
         statements = d["rule"]["features"]
 
         # the rule must start with a single logic node.
@@ -567,13 +614,19 @@ class Rule(object):
         if len(statements) != 1:
             raise InvalidRule("rule must begin with a single top level statement")
 
-        if isinstance(statements[0], capa.engine.Subscope):
+        if isinstance(statements[0], ceng.Subscope):
             raise InvalidRule("top level statement may not be a subscope")
 
         if scope not in SUPPORTED_FEATURES.keys():
             raise InvalidRule("{:s} is not a supported scope".format(scope))
 
-        return cls(name, scope, build_statements(statements[0], scope), d["rule"]["meta"], definition)
+        meta = d["rule"]["meta"]
+        if not isinstance(meta.get("att&ck", []), list):
+            raise InvalidRule("ATT&CK mapping must be a list")
+        if not isinstance(meta.get("mbc", []), list):
+            raise InvalidRule("MBC mapping must be a list")
+
+        return cls(name, scope, build_statements(statements[0], scope), meta, definition)
 
     @staticmethod
     @lru_cache()
@@ -699,7 +752,7 @@ class Rule(object):
         for key in hidden_meta.keys():
             del meta[key]
 
-        ostream = six.BytesIO()
+        ostream = io.BytesIO()
         self._get_ruamel_yaml_parser().dump(definition, ostream)
 
         for key, value in hidden_meta.items():
@@ -736,53 +789,42 @@ class Rule(object):
         # the below regex makes these adjustments and while ugly, we don't have to explore the ruamel.yaml insides
         doc = re.sub(r"!!int '0x-([0-9a-fA-F]+)'", r"-0x\1", doc)
 
+        # normalize CRLF to LF
+        doc = doc.replace("\r\n", "\n")
         return doc
 
 
-def get_rules_with_scope(rules, scope):
+def get_rules_with_scope(rules, scope) -> List[Rule]:
     """
     from the given collection of rules, select those with the given scope.
-
-    args:
-      rules (List[capa.rules.Rule]):
-      scope (str): one of the capa.rules.*_SCOPE constants.
-
-    returns:
-      List[capa.rules.Rule]:
+    `scope` is one of the capa.rules.*_SCOPE constants.
     """
     return list(rule for rule in rules if rule.scope == scope)
 
 
-def get_rules_and_dependencies(rules, rule_name):
+def get_rules_and_dependencies(rules: List[Rule], rule_name: str) -> Iterator[Rule]:
     """
     from the given collection of rules, select a rule and its dependencies (transitively).
-
-    args:
-      rules (List[Rule]):
-      rule_name (str):
-
-    yields:
-      Rule:
     """
     # we evaluate `rules` multiple times, so if its a generator, realize it into a list.
     rules = list(rules)
     namespaces = index_rules_by_namespace(rules)
-    rules = {rule.name: rule for rule in rules}
+    rules_by_name = {rule.name: rule for rule in rules}
     wanted = set([rule_name])
 
     def rec(rule):
         wanted.add(rule.name)
         for dep in rule.get_dependencies(namespaces):
-            rec(rules[dep])
+            rec(rules_by_name[dep])
 
-    rec(rules[rule_name])
+    rec(rules_by_name[rule_name])
 
-    for rule in rules.values():
+    for rule in rules_by_name.values():
         if rule.name in wanted:
             yield rule
 
 
-def ensure_rules_are_unique(rules):
+def ensure_rules_are_unique(rules: List[Rule]) -> None:
     seen = set([])
     for rule in rules:
         if rule.name in seen:
@@ -790,7 +832,7 @@ def ensure_rules_are_unique(rules):
         seen.add(rule.name)
 
 
-def ensure_rule_dependencies_are_met(rules):
+def ensure_rule_dependencies_are_met(rules: List[Rule]) -> None:
     """
     raise an exception if a rule dependency does not exist.
 
@@ -800,14 +842,14 @@ def ensure_rule_dependencies_are_met(rules):
     # we evaluate `rules` multiple times, so if its a generator, realize it into a list.
     rules = list(rules)
     namespaces = index_rules_by_namespace(rules)
-    rules = {rule.name: rule for rule in rules}
-    for rule in rules.values():
+    rules_by_name = {rule.name: rule for rule in rules}
+    for rule in rules_by_name.values():
         for dep in rule.get_dependencies(namespaces):
-            if dep not in rules:
+            if dep not in rules_by_name:
                 raise InvalidRule('rule "%s" depends on missing rule "%s"' % (rule.name, dep))
 
 
-def index_rules_by_namespace(rules):
+def index_rules_by_namespace(rules: List[Rule]) -> Dict[str, List[Rule]]:
     """
     compute the rules that fit into each namespace found within the given rules.
 
@@ -821,11 +863,6 @@ def index_rules_by_namespace(rules):
       c2/shell: [create reverse shell]
       c2/file-transfer: [download and write a file]
       c2: [create reverse shell, download and write a file]
-
-    Args:
-      rules (List[Rule]):
-
-    Returns: Dict[str, List[Rule]]
     """
     namespaces = collections.defaultdict(list)
 
@@ -841,7 +878,38 @@ def index_rules_by_namespace(rules):
     return dict(namespaces)
 
 
-class RuleSet(object):
+def topologically_order_rules(rules: List[Rule]) -> List[Rule]:
+    """
+    order the given rules such that dependencies show up before dependents.
+    this means that as we match rules, we can add features for the matches, and these
+     will be matched by subsequent rules if they follow this order.
+
+    assumes that the rule dependency graph is a DAG.
+    """
+    # we evaluate `rules` multiple times, so if its a generator, realize it into a list.
+    rules = list(rules)
+    namespaces = index_rules_by_namespace(rules)
+    rules_by_name = {rule.name: rule for rule in rules}
+    seen = set([])
+    ret = []
+
+    def rec(rule):
+        if rule.name in seen:
+            return
+
+        for dep in rule.get_dependencies(namespaces):
+            rec(rules_by_name[dep])
+
+        ret.append(rule)
+        seen.add(rule.name)
+
+    for rule in rules_by_name.values():
+        rec(rule)
+
+    return ret
+
+
+class RuleSet:
     """
     a ruleset is initialized with a collection of rules, which it verifies and sorts into scopes.
     each set of scoped rules is sorted topologically, which enables rules to match on past rule matches.
@@ -856,7 +924,7 @@ class RuleSet(object):
         capa.engine.match(ruleset.file_rules, ...)
     """
 
-    def __init__(self, rules):
+    def __init__(self, rules: List[Rule]):
         super(RuleSet, self).__init__()
 
         ensure_rules_are_unique(rules)
@@ -872,6 +940,7 @@ class RuleSet(object):
         self.function_rules = self._get_rules_for_scope(rules, FUNCTION_SCOPE)
         self.basic_block_rules = self._get_rules_for_scope(rules, BASIC_BLOCK_SCOPE)
         self.rules = {rule.name: rule for rule in rules}
+        self.rules_by_namespace = index_rules_by_namespace(rules)
 
     def __len__(self):
         return len(self.rules)
@@ -879,6 +948,9 @@ class RuleSet(object):
     def __getitem__(self, rulename):
         return self.rules[rulename]
 
+    def __contains__(self, rulename):
+        return rulename in self.rules
+
     @staticmethod
     def _get_rules_for_scope(rules, scope):
         """
@@ -899,7 +971,7 @@ class RuleSet(object):
                 continue
 
             scope_rules.update(get_rules_and_dependencies(rules, rule.name))
-        return get_rules_with_scope(capa.engine.topologically_order_rules(scope_rules), scope)
+        return get_rules_with_scope(topologically_order_rules(list(scope_rules)), scope)
 
     @staticmethod
     def _extract_subscope_rules(rules):
@@ -923,7 +995,7 @@ class RuleSet(object):
 
         return done
 
-    def filter_rules_by_meta(self, tag):
+    def filter_rules_by_meta(self, tag: str) -> "RuleSet":
         """
         return new rule set with rules filtered based on all meta field values, adds all dependency rules
         apply tag-based rule filter assuming that all required rules are loaded
@@ -932,11 +1004,11 @@ class RuleSet(object):
         TODO handle circular dependencies?
         TODO support -t=metafield <k>
         """
-        rules = self.rules.values()
+        rules = list(self.rules.values())
         rules_filtered = set([])
         for rule in rules:
             for k, v in rule.meta.items():
-                if isinstance(v, six.string_types) and tag in v:
+                if isinstance(v, str) and tag in v:
                     logger.debug('using rule "%s" and dependencies, found tag in meta.%s: %s', rule.name, k, v)
                     rules_filtered.update(set(capa.rules.get_rules_and_dependencies(rules, rule.name)))
                     break

capa/version.py

@@ -1 +1 @@
-__version__ = "1.6.0"
+__version__ = "2.0.0"

doc/installation.md

@@ -42,6 +42,7 @@ If you'd like to review and modify the capa source code, you'll need to check it
 Next, clone the capa git repository.
 We use submodules to separate [code](https://github.com/fireeye/capa), [rules](https://github.com/fireeye/capa-rules), and [test data](https://github.com/fireeye/capa-testfiles).
 To clone everything use the `--recurse-submodules` option:
+- CAUTION: The capa testfiles repository contains many malware samples. If you pull down everything using this method, you may want to install to a directory that won't trigger your anti-virus software.
 - `$ git clone --recurse-submodules https://github.com/fireeye/capa.git /local/path/to/src` (HTTPS)
 - `$ git clone --recurse-submodules git@github.com:fireeye/capa.git /local/path/to/src` (SSH)
 
@@ -59,6 +60,25 @@ Use `pip` to install the source code in "editable" mode. This means that Python
 
 You'll find that the `capa.exe` (Windows) or `capa` (Linux/MacOS) executables in your path now invoke the capa binary from this directory.
 
+#### Development
+
+##### venv [optional]
+
+For development, we recommend to use [venv](https://docs.python.org/3/tutorial/venv.html). It allows you to create a virtual environment: a self-contained directory tree that contains a Python installation for a particular version of Python, plus a number of additional packages. This approach avoids conflicts between the requirements of different applications on your computer. It also ensures that you don't overlook to add a new requirement to `setup.up` using a library already installed on your system.
+
+To create an environment (in the parent directory, to avoid commiting it by accident or messing with the linters), run:
+`$ python3 -m venv ../capa-env`
+
+To activate `capa-env` in Linux or MacOS, run:
+`$ source ../capa-env/bin/activate`
+
+To activate `capa-env` in Windows, run:
+`$ ..\capa-env\Scripts\activate.bat`
+
+For more details about creating and using virtual environments, check out the [venv documentation](https://docs.python.org/3/tutorial/venv.html).
+
+##### Install development dependencies
+
 We use the following tools to ensure consistent code style and formatting:
   - [black](https://github.com/psf/black) code formatter, with `-l 120`
   - [isort 5](https://pypi.org/project/isort/) code formatter, with `--profile black --length-sort --line-width 120`
@@ -69,30 +89,28 @@ To install these development dependencies, run:
 
 `$ pip install -e /local/path/to/src[dev]`
 
-Note that some development dependencies (including the black code formatter) require Python 3.
-
 To check the code style, formatting and run the tests you can run the script `scripts/ci.sh`.
 You can run it with the argument `no_tests` to skip the tests and only run the code style and formatting: `scripts/ci.sh no_tests`
 
+##### Setup hooks [optional]
+
+If you plan to contribute to capa, you may want to setup the hooks.
+Run `scripts/setup-hooks.sh` to set the following hooks up:
+- The `pre-commit` hook runs checks before every `git commit`.
+  It runs `scripts/ci.sh no_tests` aborting the commit if there are code style or rule linter offenses you need to fix.
+- The `pre-push` hook runs checks before every `git push`.
+  It runs `scripts/ci.sh` aborting the push if there are code style or rule linter offenses or if the tests fail.
+  This way you can ensure everything is alright before sending a pull request.
+
+You can skip the checks by using the `--no-verify` git option.
+
 ### 3. Compile binary using PyInstaller
 We compile capa standalone binaries using PyInstaller. To reproduce the build process check out the source code as described above and follow these steps.
 
 #### Install PyInstaller:
-For Python 2.7: `$ pip install 'pyinstaller==3.*'` (PyInstaller 4 doesn't support Python 2.7)
-
-For Python 3: `$ pip install 'pyinstaller`
+`$ pip install pyinstaller` (Python 3)
 
 #### Run Pyinstaller
 `$ pyinstaller .github/pyinstaller/pyinstaller.spec`
 
 You can find the compiled binary in the created directory `dist/`.
-
-### 4. Setup hooks [optional]
-If you plan to contribute to capa, you may want to setup the hooks.
-Run `scripts/setup-hooks.sh` to set the following hooks up:
-- The `pre-commit` hook runs checks before every `git commit`.
-  It runs `scripts/ci.sh no_tests` aborting the commit if there are code style or rule linter offenses you need to fix.
-  You can skip this check by using the `--no-verify` git option.
-- The `pre-push` hook runs checks before every `git push`.
-  It runs `scripts/ci.sh` aborting the push if there are code style or rule linter offenses or if the tests fail.
-  This way you can ensure everything is alright before sending a pull request.

doc/release.md

@@ -0,0 +1,43 @@
+# Release checklist
+
+- [ ] Ensure all [milestoned issues/PRs](https://github.com/fireeye/capa/milestones) are addressed, or reassign to a new milestone.
+- [ ] Add the `dont merge` label to all PRs that are close to be ready to merge (or merge them if they are ready) in [capa](https://github.com/fireeye/capa/pulls) and [capa-rules](https://github.com/fireeye/capa-rules/pulls).
+- [ ] Ensure the [CI workflow succeeds in master](https://github.com/fireeye/capa/actions/workflows/tests.yml?query=branch%3Amaster).
+- [ ] Ensure that `python scripts/lint.py rules/ --thorough` succeeds (only `missing examples`  offenses are allowed in the nursery).
+- [ ] Review changes
+  - capa https://github.com/fireeye/capa/compare/\<last-release\>...master
+  - capa-rules https://github.com/fireeye/capa-rules/compare/\<last-release>\...master
+- [ ] Update [CHANGELOG.md](https://github.com/fireeye/capa/blob/master/CHANGELOG.md)
+  - Do not forget to add a nice introduction thanking contributors
+  - Remember that we need a major release if we introduce breaking changes
+  - Sections: see template below
+  - Update `Raw diffs` links
+  - Create placeholder for `master (unreleased)` section
+    ```
+    ## master (unreleased)
+
+    ### New Features
+
+    ### Breaking Changes
+
+    ### New Rules (0)
+
+    -
+
+    ### Bug Fixes
+
+    ### capa explorer IDA Pro plugin
+
+    ### Development
+
+    ### Raw diffs
+    - [capa <release>...master](https://github.com/fireeye/capa/compare/<release>...master)
+    - [capa-rules <release>...master](https://github.com/fireeye/capa-rules/compare/<release>...master)
+    ```
+- [ ] Update [capa/version.py](https://github.com/fireeye/capa/blob/master/capa/version.py)
+- [ ] Create a PR with the updated [CHANGELOG.md](https://github.com/fireeye/capa/blob/master/CHANGELOG.md) and [capa/version.py](https://github.com/fireeye/capa/blob/master/capa/version.py). Copy this checklist in the PR description.
+- [ ] After PR review, merge the PR and [create the release in GH](https://github.com/fireeye/capa/releases/new) using text from the [CHANGELOG.md](https://github.com/fireeye/capa/blob/master/CHANGELOG.md).
+- [ ] Verify GH actions [upload artifacts](https://github.com/fireeye/capa/releases), [publish to PyPI](https://pypi.org/project/flare-capa) and [create a tag in capa rules](https://github.com/fireeye/capa-rules/tags) upon completion.
+- [ ] [Spread the word](https://twitter.com)
+- [ ] Update internal service
+

doc/usage.md

@@ -11,3 +11,9 @@ For example, `capa -t william.ballenthin@mandiant.com` runs rules that reference
 
 ### IDA Pro plugin: capa explorer
 Please check out the [capa explorer documentation](/capa/ida/plugin/README.md).
+
+### save time by reusing .viv files
+Set the environment variable `CAPA_SAVE_WORKSPACE` to instruct the underlying analysis engine to 
+cache its intermediate results to the file system. For example, vivisect will create `.viv` files.
+Subsequently, capa may run faster when reprocessing the same input file.
+This is particularly useful during rule development as you repeatedly test a rule against a known sample.

scripts/bulk-process.py

@@ -1,220 +1,223 @@
-#!/usr/bin/env python
-"""
-bulk-process
-
-Invoke capa recursively against a directory of samples
-and emit a JSON document mapping the file paths to their results.
-
-By default, this will use subprocesses for parallelism.
-Use `-n/--parallelism` to change the subprocess count from
- the default of current CPU count.
-Use `--no-mp` to use threads instead of processes,
- which is probably not useful unless you set `--parallelism=1`.
-
-example:
-
-    $ python scripts/bulk-process /tmp/suspicious
-    {
-      "/tmp/suspicious/suspicious.dll_": {
-        "rules": {
-          "encode data using XOR": {
-            "matches": {
-              "268440358": {
-              [...]
-      "/tmp/suspicious/1.dll_": { ... }
-      "/tmp/suspicious/2.dll_": { ... }
-    }
-
-
-usage:
-
-    usage: bulk-process.py [-h] [-r RULES] [-d] [-q] [-n PARALLELISM] [--no-mp]
-                           input
-
-    detect capabilities in programs.
-
-    positional arguments:
-      input                 Path to directory of files to recursively analyze
-
-    optional arguments:
-      -h, --help            show this help message and exit
-      -r RULES, --rules RULES
-                            Path to rule file or directory, use embedded rules by
-                            default
-      -d, --debug           Enable debugging output on STDERR
-      -q, --quiet           Disable all output but errors
-      -n PARALLELISM, --parallelism PARALLELISM
-                            parallelism factor
-      --no-mp               disable subprocesses
-
-Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
-Licensed under the Apache License, Version 2.0 (the "License");
- you may not use this file except in compliance with the License.
-You may obtain a copy of the License at: [package root]/LICENSE.txt
-Unless required by applicable law or agreed to in writing, software distributed under the License
- is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and limitations under the License.
-"""
-import sys
-import json
-import logging
-import os.path
-import argparse
-import multiprocessing
-import multiprocessing.pool
-
-import capa
-import capa.main
-import capa.rules
-import capa.render
-
-logger = logging.getLogger("capa")
-
-
-def get_capa_results(args):
-    """
-    run capa against the file at the given path, using the given rules.
-
-    args is a tuple, containing:
-      rules (capa.rules.RuleSet): the rules to match
-      format (str): the name of the sample file format
-      path (str): the file system path to the sample to process
-
-    args is a tuple because i'm not quite sure how to unpack multiple arguments using `map`.
-
-    returns an dict with two required keys:
-      path (str): the file system path of the sample to process
-      status (str): either "error" or "ok"
-
-    when status == "error", then a human readable message is found in property "error".
-    when status == "ok", then the capa results are found in the property "ok".
-
-    the capa results are a dictionary with the following keys:
-      meta (dict): the meta analysis results
-      capabilities (dict): the matched capabilities and their result objects
-    """
-    rules, format, path = args
-    logger.info("computing capa results for: %s", path)
-    try:
-        extractor = capa.main.get_extractor(path, format, capa.main.BACKEND_VIV, disable_progress=True)
-    except capa.main.UnsupportedFormatError:
-        # i'm 100% sure if multiprocessing will reliably raise exceptions across process boundaries.
-        # so instead, return an object with explicit success/failure status.
-        #
-        # if success, then status=ok, and results found in property "ok"
-        # if error, then status=error, and human readable message in property "error"
-        return {
-            "path": path,
-            "status": "error",
-            "error": "input file does not appear to be a PE file: %s" % path,
-        }
-    except capa.main.UnsupportedRuntimeError:
-        return {
-            "path": path,
-            "status": "error",
-            "error": "unsupported runtime or Python interpreter",
-        }
-    except Exception as e:
-        return {
-            "path": path,
-            "status": "error",
-            "error": "unexpected error: %s" % (e),
-        }
-
-    meta = capa.main.collect_metadata("", path, "", format, extractor)
-    capabilities, counts = capa.main.find_capabilities(rules, extractor, disable_progress=True)
-    meta["analysis"].update(counts)
-
-    return {
-        "path": path,
-        "status": "ok",
-        "ok": {
-            "meta": meta,
-            "capabilities": capabilities,
-        },
-    }
-
-
-def main(argv=None):
-    if argv is None:
-        argv = sys.argv[1:]
-
-        parser = argparse.ArgumentParser(description="detect capabilities in programs.")
-        capa.main.install_common_args(parser, wanted={"rules"})
-        parser.add_argument("input", type=str, help="Path to directory of files to recursively analyze")
-        parser.add_argument(
-            "-n", "--parallelism", type=int, default=multiprocessing.cpu_count(), help="parallelism factor"
-        )
-        parser.add_argument("--no-mp", action="store_true", help="disable subprocesses")
-        args = parser.parse_args(args=argv)
-        capa.main.handle_common_args(args)
-
-        if args.rules == "(embedded rules)":
-            logger.info("using default embedded rules")
-            logger.debug("detected running from source")
-            args.rules = os.path.join(os.path.dirname(__file__), "..", "rules")
-            logger.debug("default rule path (source method): %s", args.rules)
-        else:
-            logger.info("using rules path: %s", args.rules)
-
-        try:
-            rules = capa.main.get_rules(args.rules)
-            rules = capa.rules.RuleSet(rules)
-            logger.info("successfully loaded %s rules", len(rules))
-        except (IOError, capa.rules.InvalidRule, capa.rules.InvalidRuleSet) as e:
-            logger.error("%s", str(e))
-            return -1
-
-        samples = []
-        for (base, directories, files) in os.walk(args.input):
-            for file in files:
-                samples.append(os.path.join(base, file))
-
-        def pmap(f, args, parallelism=multiprocessing.cpu_count()):
-            """apply the given function f to the given args using subprocesses"""
-            return multiprocessing.Pool(parallelism).imap(f, args)
-
-        def tmap(f, args, parallelism=multiprocessing.cpu_count()):
-            """apply the given function f to the given args using threads"""
-            return multiprocessing.pool.ThreadPool(parallelism).imap(f, args)
-
-        def map(f, args, parallelism=None):
-            """apply the given function f to the given args in the current thread"""
-            for arg in args:
-                yield f(arg)
-
-        if args.no_mp:
-            if args.parallelism == 1:
-                logger.debug("using current thread mapper")
-                mapper = map
-            else:
-                logger.debug("using threading mapper")
-                mapper = tmap
-        else:
-            logger.debug("using process mapper")
-            mapper = pmap
-
-        results = {}
-        for result in mapper(
-            get_capa_results, [(rules, "pe", sample) for sample in samples], parallelism=args.parallelism
-        ):
-            if result["status"] == "error":
-                logger.warning(result["error"])
-            elif result["status"] == "ok":
-                meta = result["ok"]["meta"]
-                capabilities = result["ok"]["capabilities"]
-                # our renderer expects to emit a json document for a single sample
-                # so we deserialize the json document, store it in a larger dict, and we'll subsequently re-encode.
-                results[result["path"]] = json.loads(capa.render.render_json(meta, rules, capabilities))
-            else:
-                raise ValueError("unexpected status: %s" % (result["status"]))
-
-        print(json.dumps(results))
-
-        logger.info("done.")
-
-        return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())
+#!/usr/bin/env python
+"""
+bulk-process
+
+Invoke capa recursively against a directory of samples
+and emit a JSON document mapping the file paths to their results.
+
+By default, this will use subprocesses for parallelism.
+Use `-n/--parallelism` to change the subprocess count from
+ the default of current CPU count.
+Use `--no-mp` to use threads instead of processes,
+ which is probably not useful unless you set `--parallelism=1`.
+
+example:
+
+    $ python scripts/bulk-process /tmp/suspicious
+    {
+      "/tmp/suspicious/suspicious.dll_": {
+        "rules": {
+          "encode data using XOR": {
+            "matches": {
+              "268440358": {
+              [...]
+      "/tmp/suspicious/1.dll_": { ... }
+      "/tmp/suspicious/2.dll_": { ... }
+    }
+
+
+usage:
+
+    usage: bulk-process.py [-h] [-r RULES] [-d] [-q] [-n PARALLELISM] [--no-mp]
+                           input
+
+    detect capabilities in programs.
+
+    positional arguments:
+      input                 Path to directory of files to recursively analyze
+
+    optional arguments:
+      -h, --help            show this help message and exit
+      -r RULES, --rules RULES
+                            Path to rule file or directory, use embedded rules by
+                            default
+      -d, --debug           Enable debugging output on STDERR
+      -q, --quiet           Disable all output but errors
+      -n PARALLELISM, --parallelism PARALLELISM
+                            parallelism factor
+      --no-mp               disable subprocesses
+
+Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+You may obtain a copy of the License at: [package root]/LICENSE.txt
+Unless required by applicable law or agreed to in writing, software distributed under the License
+ is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and limitations under the License.
+"""
+import os
+import sys
+import json
+import logging
+import os.path
+import argparse
+import multiprocessing
+import multiprocessing.pool
+
+import capa
+import capa.main
+import capa.rules
+import capa.render.json
+
+logger = logging.getLogger("capa")
+
+
+def get_capa_results(args):
+    """
+    run capa against the file at the given path, using the given rules.
+
+    args is a tuple, containing:
+      rules (capa.rules.RuleSet): the rules to match
+      signatures (List[str]): list of file system paths to signature files
+      format (str): the name of the sample file format
+      path (str): the file system path to the sample to process
+
+    args is a tuple because i'm not quite sure how to unpack multiple arguments using `map`.
+
+    returns an dict with two required keys:
+      path (str): the file system path of the sample to process
+      status (str): either "error" or "ok"
+
+    when status == "error", then a human readable message is found in property "error".
+    when status == "ok", then the capa results are found in the property "ok".
+
+    the capa results are a dictionary with the following keys:
+      meta (dict): the meta analysis results
+      capabilities (dict): the matched capabilities and their result objects
+    """
+    rules, sigpaths, format, path = args
+    should_save_workspace = os.environ.get("CAPA_SAVE_WORKSPACE") not in ("0", "no", "NO", "n", None)
+    logger.info("computing capa results for: %s", path)
+    try:
+        extractor = capa.main.get_extractor(
+            path, format, capa.main.BACKEND_VIV, sigpaths, should_save_workspace, disable_progress=True
+        )
+    except capa.main.UnsupportedFormatError:
+        # i'm 100% sure if multiprocessing will reliably raise exceptions across process boundaries.
+        # so instead, return an object with explicit success/failure status.
+        #
+        # if success, then status=ok, and results found in property "ok"
+        # if error, then status=error, and human readable message in property "error"
+        return {
+            "path": path,
+            "status": "error",
+            "error": "input file does not appear to be a PE file: %s" % path,
+        }
+    except capa.main.UnsupportedRuntimeError:
+        return {
+            "path": path,
+            "status": "error",
+            "error": "unsupported runtime or Python interpreter",
+        }
+    except Exception as e:
+        return {
+            "path": path,
+            "status": "error",
+            "error": "unexpected error: %s" % (e),
+        }
+
+    meta = capa.main.collect_metadata("", path, "", format, extractor)
+    capabilities, counts = capa.main.find_capabilities(rules, extractor, disable_progress=True)
+    meta["analysis"].update(counts)
+
+    return {
+        "path": path,
+        "status": "ok",
+        "ok": {
+            "meta": meta,
+            "capabilities": capabilities,
+        },
+    }
+
+
+def main(argv=None):
+    if argv is None:
+        argv = sys.argv[1:]
+
+        parser = argparse.ArgumentParser(description="detect capabilities in programs.")
+        capa.main.install_common_args(parser, wanted={"rules", "signatures"})
+        parser.add_argument("input", type=str, help="Path to directory of files to recursively analyze")
+        parser.add_argument(
+            "-n", "--parallelism", type=int, default=multiprocessing.cpu_count(), help="parallelism factor"
+        )
+        parser.add_argument("--no-mp", action="store_true", help="disable subprocesses")
+        args = parser.parse_args(args=argv)
+        capa.main.handle_common_args(args)
+
+        try:
+            rules = capa.main.get_rules(args.rules)
+            rules = capa.rules.RuleSet(rules)
+            logger.info("successfully loaded %s rules", len(rules))
+        except (IOError, capa.rules.InvalidRule, capa.rules.InvalidRuleSet) as e:
+            logger.error("%s", str(e))
+            return -1
+
+        try:
+            sig_paths = capa.main.get_signatures(args.signatures)
+        except (IOError) as e:
+            logger.error("%s", str(e))
+            return -1
+
+        samples = []
+        for (base, directories, files) in os.walk(args.input):
+            for file in files:
+                samples.append(os.path.join(base, file))
+
+        def pmap(f, args, parallelism=multiprocessing.cpu_count()):
+            """apply the given function f to the given args using subprocesses"""
+            return multiprocessing.Pool(parallelism).imap(f, args)
+
+        def tmap(f, args, parallelism=multiprocessing.cpu_count()):
+            """apply the given function f to the given args using threads"""
+            return multiprocessing.pool.ThreadPool(parallelism).imap(f, args)
+
+        def map(f, args, parallelism=None):
+            """apply the given function f to the given args in the current thread"""
+            for arg in args:
+                yield f(arg)
+
+        if args.no_mp:
+            if args.parallelism == 1:
+                logger.debug("using current thread mapper")
+                mapper = map
+            else:
+                logger.debug("using threading mapper")
+                mapper = tmap
+        else:
+            logger.debug("using process mapper")
+            mapper = pmap
+
+        results = {}
+        for result in mapper(
+            get_capa_results, [(rules, sig_paths, "pe", sample) for sample in samples], parallelism=args.parallelism
+        ):
+            if result["status"] == "error":
+                logger.warning(result["error"])
+            elif result["status"] == "ok":
+                meta = result["ok"]["meta"]
+                capabilities = result["ok"]["capabilities"]
+                # our renderer expects to emit a json document for a single sample
+                # so we deserialize the json document, store it in a larger dict, and we'll subsequently re-encode.
+                results[result["path"]] = json.loads(capa.render.json.render(meta, rules, capabilities))
+            else:
+                raise ValueError("unexpected status: %s" % (result["status"]))
+
+        print(json.dumps(results))
+
+        logger.info("done.")
+
+        return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

scripts/capa2yara.py

@@ -0,0 +1,760 @@
+"""
+Convert capa rules to YARA rules (where this is possible)
+
+    usage: capa2yara.py [-h] [--private] [--version] [-v] [-vv] [-d] [-q] [--color {auto,always,never}] [-t TAG] rules
+
+Capa to YARA rule converter
+
+positional arguments:
+  rules                 Path to rules
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --private, -p         Create private rules
+  --version             show program's version number and exit
+  -v, --verbose         enable verbose result document (no effect with --json)
+  -vv, --vverbose       enable very verbose result document (no effect with --json)
+  -d, --debug           enable debugging output on STDERR
+  -q, --quiet           disable all output but errors
+  --color {auto,always,never}
+                        enable ANSI color codes in results, default: only during interactive session
+  -t TAG, --tag TAG     filter on rule meta field values
+
+
+Copyright (C) 2020, 2021 Arnim Rupp (@ruppde) and FireEye, Inc. All Rights Reserved.
+Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+You may obtain a copy of the License at: [package root]/LICENSE.txt
+Unless required by applicable law or agreed to in writing, software distributed under the License
+ is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and limitations under the License.
+"""
+
+import re
+import sys
+import string
+import logging
+import argparse
+import datetime
+import itertools
+
+import capa.main
+import capa.rules
+import capa.engine
+import capa.features
+import capa.features.insn
+from capa.features.common import ARCH_X32, ARCH_X64, String
+
+logger = logging.getLogger("capa2yara")
+
+today = str(datetime.date.today())
+
+# create unique variable names for each rule in case somebody wants to move/copy stuff around later
+var_names = ["".join(letters) for letters in itertools.product(string.ascii_lowercase, repeat=3)]
+
+
+# this have to be the internal names used by capa.py which are sometimes different to the ones written out in the rules, e.g. "2 or more" is "Some", count is Range
+unsupported = ["characteristic", "mnemonic", "offset", "subscope", "Range"]
+# TODO shorten this list, possible stuff:
+# - 2 or more strings: e.g.
+# -- https://github.com/fireeye/capa-rules/blob/master/collection/file-managers/gather-direct-ftp-information.yml
+# -- https://github.com/fireeye/capa-rules/blob/master/collection/browser/gather-firefox-profile-information.yml
+# - count(string    (1 rule: /executable/subfile/pe/contain-an-embedded-pe-file.yml)
+# - count(match( could be done by creating the referenced rule a 2nd time with the condition, that it hits x times (only 1 rule: ./anti-analysis/anti-disasm/contain-anti-disasm-techniques.yml)
+# - it would be technically possible to get the "basic blocks" working, but the rules contain mostly other non supported statements in there => not worth the effort.
+
+# collect all converted rules to be able to check if we have needed sub rules for match:
+converted_rules = []
+count_incomplete = 0
+
+default_tags = "CAPA "
+
+# minimum number of rounds to do be able to convert rules which depend on referenced rules in several levels of depth
+min_rounds = 5
+
+unsupported_capa_rules = open("unsupported_capa_rules.yml", "wb")
+unsupported_capa_rules_names = open("unsupported_capa_rules.txt", "wb")
+unsupported_capa_rules_list = []
+
+condition_header = """
+    capa_pe_file and
+"""
+
+condition_rule = """
+private rule capa_pe_file : CAPA {
+    meta:
+        description = "match in PE files. used by all further CAPA rules"
+        author = "Arnim Rupp"
+    condition:
+        uint16be(0) == 0x4d5a
+        or uint16be(0) == 0x558b
+        or uint16be(0) == 0x5649
+}
+"""
+
+
+def check_feature(statement, rulename):
+    if statement in unsupported:
+        logger.info("unsupported: " + statement + " in rule: " + rulename)
+        return True
+    else:
+        return False
+
+
+def get_rule_url(path):
+    path = re.sub(r"\.\.\/", "", path)
+    path = re.sub(r"capa-rules\/", "", path)
+    return "https://github.com/fireeye/capa-rules/blob/master/" + path
+
+
+def convert_capa_number_to_yara_bytes(number):
+    if not number.startswith("0x"):
+        print("TODO: fix decimal")
+        sys.exit()
+
+    number = re.sub(r"^0[xX]", "", number)
+    logger.info("number ok: " + repr(number))
+
+    # include spaces every 2 hex
+    bytesv = re.sub(r"(..)", r"\1 ", number)
+
+    # reverse order
+    bytesl = bytesv.split(" ")
+    bytesl.reverse()
+    bytesv = " ".join(bytesl)
+
+    # fix spaces
+    bytesv = bytesv[1:] + " "
+
+    return bytesv
+
+
+def convert_rule_name(rule_name):
+
+    # yara rule names: "Identifiers must follow the same lexical conventions of the C programming language, they can contain any alphanumeric character and the underscore character, but the first character cannot be a digit. Rule identifiers are case sensitive and cannot exceed 128 characters." so we replace any non-alpanum with _
+    rule_name = re.sub(r"\W", "_", rule_name)
+    rule_name = "capa_" + rule_name
+
+    return rule_name
+
+
+def convert_description(statement):
+    try:
+        desc = statement.description
+        if desc:
+            yara_desc = " // " + desc
+            logger.info("using desc: " + repr(yara_desc))
+            return yara_desc
+    except:
+        # no description
+        pass
+
+    return ""
+
+
+def convert_rule(rule, rulename, cround, depth):
+
+    depth += 1
+    logger.info("recursion depth: " + str(depth))
+
+    global var_names
+
+    def do_statement(s_type, kid):
+        yara_strings = ""
+        yara_condition = ""
+        if check_feature(s_type, rulename):
+            return "BREAK", s_type
+        elif s_type == "string":
+            string = kid.value
+            logger.info("doing string: " + repr(string))
+            string = string.replace("\\", "\\\\")
+            string = string.replace("\n", "\\n")
+            string = string.replace("\t", "\\t")
+            var_name = "str_" + var_names.pop(0)
+            yara_strings += "\t$" + var_name + ' = "' + string + '" ascii wide' + convert_description(kid) + "\n"
+            yara_condition += "\t$" + var_name + " "
+        elif s_type == "api" or s_type == "import":
+            # TODO: is it possible in YARA to make a difference between api & import?
+
+            # https://github.com/fireeye/capa-rules/blob/master/doc/format.md#api
+            api = kid.value
+            logger.info("doing api: " + repr(api))
+
+            #    e.g. kernel32.CreateNamedPipe => look for kernel32.dll and CreateNamedPipe
+            if "." in api:
+                dll, api = api.split(".")
+
+                # usage of regex is needed and /i because string search for "CreateMutex" in imports() doesn't look for e.g. CreateMutexA
+                yara_condition += "\tpe.imports(/" + dll + "/i, /" + api + "/) "
+
+            else:
+                # e.g. - api: 'CallNextHookEx'
+                # (from user32.dll)
+
+                # even looking for empty string in dll_regex doesn't work for some files (list below) with pe.imports so do just a string search
+                # yara_condition += '\tpe.imports(/.{0,30}/i, /' + api + '/) '
+                # 5fbbfeed28b258c42e0cfeb16718b31c, 2D3EDC218A90F03089CC01715A9F047F, 7EFF498DE13CC734262F87E6B3EF38AB, C91887D861D9BD4A5872249B641BC9F9, a70052c45e907820187c7e6bcdc7ecca, 0596C4EA5AA8DEF47F22C85D75AACA95
+                var_name = "api_" + var_names.pop(0)
+
+                # limit regex with word boundary \b but also search for appended A and W
+                # TODO: better use something like /(\\x00|\\x01|\\x02|\\x03|\\x04)' + api + '(A|W)?\\x00/  ???
+                yara_strings += "\t$" + var_name + " = /\\b" + api + "(A|W)?\\b/ ascii wide\n"
+                yara_condition += "\t$" + var_name + " "
+
+        elif s_type == "export":
+            export = kid.value
+            logger.info("doing export: " + repr(export))
+
+            yara_condition += '\tpe.exports("' + export + '") '
+
+        elif s_type == "section":
+            # https://github.com/fireeye/capa-rules/blob/master/doc/format.md#section
+            section = kid.value
+            logger.info("doing section: " + repr(section))
+
+            # e.g. - section: .rsrc
+            var_name_sec = var_names.pop(0)
+            # yeah, it would be better to make one loop out of multiple sections but we're in POC-land (and I guess it's not much of a performance hit, loop over short array?)
+            yara_condition += (
+                "\tfor any " + var_name_sec + " in pe.sections : ( " + var_name_sec + '.name == "' + section + '" ) '
+            )
+
+        elif s_type == "match":
+            # https://github.com/fireeye/capa-rules/blob/master/doc/format.md#matching-prior-rule-matches-and-namespaces
+            match = kid.value
+            logger.info("doing match: " + repr(match))
+
+            # e.g. - match: create process
+            #      - match: host-interaction/file-system/write
+            match_rule_name = convert_rule_name(match)
+
+            if match.startswith(rulename + "/"):
+                logger.info("Depending on myself = basic block: " + match)
+                return "BREAK", "Depending on myself = basic block"
+
+            if match_rule_name in converted_rules:
+                yara_condition += "\t" + match_rule_name + "\n"
+            else:
+                # don't complain in the early rounds as there should be 3+ rounds (if all rules are converted)
+                if cround > min_rounds - 2:
+                    logger.info("needed sub-rule not converted (yet, maybe in next round): " + repr(match))
+                    return "BREAK", "needed sub-rule not converted"
+                else:
+                    return "BREAK", "NOLOG"
+
+        elif s_type == "bytes":
+            bytesv = kid.get_value_str()
+            logger.info("doing bytes: " + repr(bytesv))
+            var_name = var_names.pop(0)
+
+            yara_strings += "\t$" + var_name + " = { " + bytesv + " }" + convert_description(kid) + "\n"
+            yara_condition += "\t$" + var_name + " "
+
+        elif s_type == "number":
+            number = kid.get_value_str()
+            logger.info("doing number: " + repr(number))
+
+            if len(number) < 10:
+                logger.info("too short for byte search (until I figure out how to do it properly)" + repr(number))
+                return "BREAK", "Number too short"
+
+            # there's just one rule which contains 0xFFFFFFF but yara gives a warning if if used
+            if number == "0xFFFFFFFF":
+                return "BREAK", "slow byte pattern for YARA search"
+
+            logger.info("number ok: " + repr(number))
+            number = convert_capa_number_to_yara_bytes(number)
+            logger.info("number ok: " + repr(number))
+
+            var_name = "num_" + var_names.pop(0)
+            yara_strings += "\t$" + var_name + " = { " + number + "}" + convert_description(kid) + "\n"
+            yara_condition += "$" + var_name + " "
+
+        elif s_type == "regex":
+            regex = kid.get_value_str()
+            logger.info("doing regex: " + repr(regex))
+
+            # change capas /xxx/i to yaras /xxx/ nocase, count will be used later to decide appending 'nocase'
+            regex, count = re.subn(r"/i$", "/", regex)
+
+            # remove / in the begining and end
+            regex = regex[1:-1]
+
+            # all .* in the regexes of capa look like they should be maximum 100 chars so take 1000 to speed up rules and prevent yara warnings on poor performance
+            regex = regex.replace(".*", ".{,1000}")
+            # strange: capa accepts regexes with unsescaped / like - string: /com/exe4j/runtime/exe4jcontroller/i in capa-rules/compiler/exe4j/compiled-with-exe4j.yml, needs a fix for yara:
+            # would assume that get_value_str() gives the raw string
+            regex = re.sub(r"(?<!\\)/", r"\/", regex)
+
+            # capa uses python regex which accepts /reg(|.exe)/ but yaras regex engine doesn't not => fix it
+            # /reg(|.exe)/ => /reg(.exe)?/
+            regex = re.sub(r"\(\|([^\)]+)\)", r"(\1)?", regex)
+
+            # change begining of line to null byte, e.g. /^open => /\x00open (not word boundary because we're not looking for the begining of a word in a text but usually a function name if there's ^ in a capa rule)
+            regex = re.sub(r"^\^", r"\\x00", regex)
+
+            # regex = re.sub(r"^\^", r"\\b", regex)
+
+            regex = "/" + regex + "/"
+            if count:
+                regex += " nocase"
+
+            # strange: if statement.name == "string", the string is as it is, if statement.name == "regex", the string has // around it, e.g. /regex/
+            var_name = "re_" + var_names.pop(0)
+            yara_strings += "\t" + "$" + var_name + " = " + regex + " ascii wide " + convert_description(kid) + "\n"
+            yara_condition += "\t" + "$" + var_name + " "
+        elif s_type == "Not" or s_type == "And" or s_type == "Or":
+            pass
+        else:
+            logger.info("something unhandled: " + repr(s_type))
+            sys.exit()
+
+        return yara_strings, yara_condition
+
+    ############################## end def do_statement
+
+    yara_strings_list = []
+    yara_condition_list = []
+    rule_comment = ""
+    incomplete = 0
+
+    statement = rule.name
+
+    logger.info("doing statement: " + statement)
+
+    if check_feature(statement, rulename):
+        return "BREAK", statement, rule_comment, incomplete
+
+    if statement == "And" or statement == "Or":
+        desc = convert_description(rule)
+        if desc:
+            logger.info("description of bool statement: " + repr(desc))
+            yara_strings_list.append("\t" * depth + desc + "\n")
+    elif statement == "Not":
+        logger.info("one of those seldom nots: " + rule.name)
+
+    # check for nested statements
+    try:
+        kids = rule.children
+        num_kids = len(kids)
+        logger.info("kids: " + kids)
+    except:
+        logger.info("no kids in rule: " + rule.name)
+
+        try:
+            # maybe it's "Not" = only one child:
+            kid = rule.child
+            kids = [kid]
+            num_kids = 1
+            logger.info("kid: %s", kids)
+        except:
+            logger.info("no kid in rule: %s", rule.name)
+
+    # just a single statement without 'and' or 'or' before it in this rule
+    if "kids" not in locals().keys():
+        logger.info("no kids: " + rule.name)
+
+        yara_strings_sub, yara_condition_sub = do_statement(statement, rule)
+
+        if yara_strings_sub == "BREAK":
+            logger.info("Unknown feature at1: " + rule.name)
+            return "BREAK", yara_condition_sub, rule_comment, incomplete
+        yara_strings_list.append(yara_strings_sub)
+        yara_condition_list.append(yara_condition_sub)
+
+    else:
+        x = 0
+        logger.info("doing kids: %r - len: %s", kids, num_kids)
+        for kid in kids:
+            s_type = kid.name
+            logger.info("doing type: " + s_type + " kidnum: " + str(x))
+
+            if s_type == "Some":
+                cmin = kid.count
+                logger.info("Some type with mininum: " + str(cmin))
+
+                if not cmin:
+                    logger.info("this is optional: which means, we can just ignore it")
+                    x += 1
+                    continue
+                elif statement == "Or":
+                    logger.info("we're inside an OR, we can just ignore it")
+                    x += 1
+                    continue
+                else:
+                    # this is "x or more". could be coded for strings TODO
+                    return "BREAK", "Some aka x or more (TODO)", rule_comment, incomplete
+
+            if s_type == "And" or s_type == "Or" or s_type == "Not" and not kid.name == "Some":
+                logger.info("doing bool with recursion: " + repr(kid))
+                logger.info("kid coming: " + repr(kid.name))
+                # logger.info("grandchildren: " + repr(kid.children))
+
+                ##### here we go into RECURSION  ##################################################################################
+                yara_strings_sub, yara_condition_sub, rule_comment_sub, incomplete_sub = convert_rule(
+                    kid, rulename, cround, depth
+                )
+
+                logger.info("coming out of this recursion, depth: " + repr(depth) + " s_type: " + s_type)
+
+                if yara_strings_sub == "BREAK":
+                    logger.info(
+                        "Unknown feature at2: " + rule.name + " - s_type: " + s_type + " - depth: " + str(depth)
+                    )
+
+                    # luckily this is only a killer, if we're inside an 'And', inside 'Or' we're just missing some coverage
+                    # only accept incomplete rules in rounds > 3 because the reason might be a reference to another rule not converted yet because of missing dependencies
+                    logger.info("rule.name,  depth,  cround: " + rule.name + ", " + str(depth) + ", " + str(cround))
+                    if rule.name == "Or" and depth == 1 and cround > min_rounds - 1:
+                        logger.info(
+                            "Unknown feature, just ignore this branch and keep the rest bec we're in Or (1): "
+                            + s_type
+                            + " - depth: "
+                            + str(depth)
+                        )
+                        # remove last 'or'
+                        # yara_condition = re.sub(r'\sor $', ' ', yara_condition)
+                        rule_comment += "This rule is incomplete because a branch inside an Or-statement had an unsupported feature and was skipped => coverage is reduced compared to the original capa rule. "
+                        x += 1
+                        incomplete = 1
+                        continue
+                    else:
+                        return "BREAK", yara_condition_sub, rule_comment, incomplete
+
+                rule_comment += rule_comment_sub
+                yara_strings_list.append(yara_strings_sub)
+                yara_condition_list.append(yara_condition_sub)
+
+                incomplete = incomplete or incomplete_sub
+
+            yara_strings_sub, yara_condition_sub = do_statement(s_type, kid)
+
+            if yara_strings_sub == "BREAK":
+                logger.info("Unknown feature at3: " + rule.name)
+                logger.info("rule.name,  depth,  cround: " + rule.name + ", " + str(depth) + ", " + str(cround))
+                if rule.name == "Or" and depth == 1 and cround > min_rounds - 1:
+                    logger.info(
+                        "Unknown feature, just ignore this branch and keep the rest bec we're in Or (2): "
+                        + s_type
+                        + " - depth: "
+                        + str(depth)
+                    )
+
+                    rule_comment += "This rule is incomplete because a branch inside an Or-statement had an unsupported feature and was skipped => coverage is reduced compared to the original capa rule. "
+                    x += 1
+                    incomplete = 1
+                    continue
+                else:
+                    return "BREAK", yara_condition_sub, rule_comment, incomplete
+
+            # don't append And or Or if we got no condition back from this kid from e.g. match in myself or unsupported feature inside 'Or'
+            if not yara_condition_sub:
+                continue
+
+            yara_strings_list.append(yara_strings_sub)
+            yara_condition_list.append(yara_condition_sub)
+            x += 1
+
+    # this might happen, if all conditions are inside "or" and none of them was supported
+    if not yara_condition_list:
+        return (
+            "BREAK",
+            'Multiple statements inside "- or:" where all unsupported, the last one was "' + s_type + '"',
+            rule_comment,
+            incomplete,
+        )
+
+    if statement == "And" or statement == "Or":
+        if yara_strings_list:
+            yara_strings = "".join(yara_strings_list)
+        else:
+            yara_strings = ""
+
+        yara_condition = " (\n\t\t" + ("\n\t\t" + statement.lower() + " ").join(yara_condition_list) + " \n\t) "
+
+    elif statement == "Some":
+        cmin = rule.count
+        logger.info("Some type with mininum at2: " + str(cmin))
+
+        if not cmin:
+            logger.info("this is optional: which means, we can just ignore it")
+        else:
+            # this is "x or more". could be coded for strings TODO
+            return "BREAK", "Some aka x or more (TODO)", rule_comment, incomplete
+    elif statement == "Not":
+        logger.info("Not")
+        yara_strings = "".join(yara_strings_list)
+        yara_condition = "not " + "".join(yara_condition_list) + " "
+    else:
+        if len(yara_condition_list) != 1:
+            logger.info("something wrong around here" + repr(yara_condition_list) + " - " + statement)
+            sys.exit()
+
+        # strings might be empty with only conditions
+        if yara_strings_list:
+            yara_strings = "\n\t" + yara_strings_list[0]
+
+        yara_condition = "\n\t" + yara_condition_list[0]
+
+    logger.info(
+        f"################# end of convert_rule() #strings: {len(yara_strings_list)} #conditions: {len(yara_condition_list)}"
+    )
+    logger.info(f"strings: {yara_strings} conditions: {yara_condition}")
+
+    return yara_strings, yara_condition, rule_comment, incomplete
+
+
+def output_yar(yara):
+    print(yara + "\n")
+
+
+def output_unsupported_capa_rules(yaml, capa_rulename, url, reason):
+
+    if reason != "NOLOG":
+        if capa_rulename not in unsupported_capa_rules_list:
+            logger.info("unsupported: " + capa_rulename + " - reason: " + reason + " - url: " + url)
+
+            unsupported_capa_rules_list.append(capa_rulename)
+            unsupported_capa_rules.write(yaml.encode("utf-8") + b"\n")
+            unsupported_capa_rules.write(
+                (
+                    "Reason: "
+                    + reason
+                    + " (there might be multiple unsupported things in this rule, this is the 1st one encountered)"
+                ).encode("utf-8")
+                + b"\n"
+            )
+            unsupported_capa_rules.write(url.encode("utf-8") + b"\n----------------------------------------------\n")
+            unsupported_capa_rules_names.write(capa_rulename.encode("utf-8") + b":")
+            unsupported_capa_rules_names.write(reason.encode("utf-8") + b":")
+            unsupported_capa_rules_names.write(url.encode("utf-8") + b"\n")
+
+
+def convert_rules(rules, namespaces, cround):
+    for rule in rules.rules.values():
+
+        rule_name = convert_rule_name(rule.name)
+
+        if rule.meta.get("capa/subscope-rule", False):
+            logger.info("skipping sub scope rule capa: " + rule.name)
+            continue
+
+        if rule_name in converted_rules:
+            logger.info("skipping already converted rule capa: " + rule.name + " - yara rule: " + rule_name)
+            continue
+
+        logger.info("-------------------------- DOING RULE CAPA: " + rule.name + " - yara rule: " + rule_name)
+        if "capa/path" in rule.meta:
+            url = get_rule_url(rule.meta["capa/path"])
+        else:
+            url = "no url"
+
+        logger.info("URL: " + url)
+        logger.info("statements: " + repr(rule.statement))
+
+        # don't really know what that passed empty string is good for :)
+        dependencies = rule.get_dependencies(namespaces)
+
+        if len(dependencies):
+            logger.info("Dependencies at4: " + rule.name + " - dep: " + str(dependencies))
+
+            for dep in dependencies:
+                logger.info("Dependencies at44: " + dep)
+                if not dep.startswith(rule.name + "/"):
+                    logger.info("Depending on another rule: " + dep)
+                    continue
+
+        yara_strings, yara_condition, rule_comment, incomplete = convert_rule(rule.statement, rule.name, cround, 0)
+
+        if yara_strings == "BREAK":
+            # only give up if in final extra round #9000
+            if cround == 9000:
+                output_unsupported_capa_rules(rule.to_yaml(), rule.name, url, yara_condition)
+            logger.info("Unknown feature at5: " + rule.name)
+        else:
+
+            yara_meta = ""
+            metas = rule.meta
+            rule_tags = ""
+
+            for meta in metas:
+                meta_name = meta
+                # e.g. 'examples:' can be a list
+                seen_hashes = []
+                if isinstance(metas[meta], list):
+                    if meta_name == "examples":
+                        meta_name = "hash"
+                    if meta_name == "att&ck":
+                        meta_name = "attack"
+                        for attack in list(metas[meta]):
+                            logger.info("attack:" + attack)
+                            # cut out tag in square brackets, e.g. Defense Evasion::Obfuscated Files or Information [T1027] => T1027
+                            r = re.search(r"\[(T[^\]]*)", attack)
+                            if r:
+                                tag = r.group(1)
+                                logger.info("attack tag:" + tag)
+                                tag = re.sub(r"\W", "_", tag)
+                                rule_tags += tag + " "
+                                # also add a line "attack = ..." to yaras 'meta:' to keep the long description:
+                                yara_meta += '\tattack = "' + attack + '"\n'
+                    elif meta_name == "mbc":
+                        for mbc in list(metas[meta]):
+                            logger.info("mbc:" + mbc)
+                            # cut out tag in square brackets, e.g. Cryptography::Encrypt Data::RC6 [C0027.010] => C0027.010
+                            r = re.search(r"\[(.[^\]]*)", mbc)
+                            if r:
+                                tag = r.group(1)
+                                logger.info("mbc tag:" + tag)
+                                tag = re.sub(r"\W", "_", tag)
+                                rule_tags += tag + " "
+
+                                # also add a line "mbc = ..." to yaras 'meta:' to keep the long description:
+                                yara_meta += '\tmbc = "' + mbc + '"\n'
+
+                    for value in metas[meta]:
+                        if meta_name == "hash":
+                            value = re.sub(r"^([0-9a-f]{20,64}):0x[0-9a-f]{1,10}$", r"\1", value, flags=re.IGNORECASE)
+
+                            # examples in capa can contain the same hash several times with different offset, so check if it's already there:
+                            # (keeping the offset might be interessting for some but breaks yara-ci for checking of the final rules
+                            if not value in seen_hashes:
+                                yara_meta += "\t" + meta_name + ' = "' + value + '"\n'
+                                seen_hashes.append(value)
+
+                else:
+                    # no list:
+                    if meta == "capa/path":
+                        url = get_rule_url(metas[meta])
+                        meta_name = "reference"
+                        meta_value = "This YARA rule converted from capa rule: " + url
+                    else:
+                        meta_value = metas[meta]
+
+                    if meta_name == "name":
+                        meta_name = "description"
+                        meta_value += " (converted from capa rule)"
+                    elif meta_name == "lib":
+                        meta_value = str(meta_value)
+                    elif meta_name == "capa/nursery":
+                        meta_name = "capa_nursery"
+                        meta_value = str(meta_value)
+
+                    # for the rest of the maec/malware-category names:
+                    meta_name = re.sub(r"\W", "_", meta_name)
+
+                    if meta_name and meta_value:
+                        yara_meta += "\t" + meta_name + ' = "' + meta_value + '"\n'
+
+            rule_name_bonus = ""
+            if rule_comment:
+                yara_meta += '\tcomment = "' + rule_comment + '"\n'
+            yara_meta += '\tdate = "' + today + '"\n'
+            yara_meta += '\tminimum_yara = "3.8"\n'
+            yara_meta += '\tlicense = "Apache-2.0 License"\n'
+
+            # check if there's some beef in condition:
+            tmp_yc = re.sub(r"(and|or|not)", "", yara_condition)
+            if re.search(r"\w", tmp_yc):
+
+                yara = ""
+                if make_priv:
+                    yara = "private "
+
+                # put yara rule tags here:
+                rule_tags = default_tags + rule_tags
+                yara += "rule " + rule_name + " : " + rule_tags + " { \n  meta: \n " + yara_meta + "\n"
+
+                if "$" in yara_strings:
+                    yara += "  strings: \n " + yara_strings + " \n"
+
+                yara += "  condition:" + condition_header + yara_condition + "\n}"
+
+                # TODO: now the rule is finished and could be automatically checked with the capa-testfile(s) named in meta (doing it for all of them using yara-ci upload at the moment)
+                output_yar(yara)
+                converted_rules.append(rule_name)
+                global count_incomplete
+                count_incomplete += incomplete
+            else:
+                output_unsupported_capa_rules(rule.to_yaml(), rule.name, url, yara_condition)
+                pass
+
+
+def main(argv=None):
+    if argv is None:
+        argv = sys.argv[1:]
+
+    parser = argparse.ArgumentParser(description="Capa to YARA rule converter")
+    parser.add_argument("rules", type=str, help="Path to rules")
+    parser.add_argument("--private", "-p", action="store_true", help="Create private rules", default=False)
+    capa.main.install_common_args(parser, wanted={"tag"})
+
+    args = parser.parse_args(args=argv)
+    global make_priv
+    make_priv = args.private
+
+    if args.verbose:
+        level = logging.DEBUG
+    elif args.quiet:
+        level = logging.ERROR
+    else:
+        level = logging.INFO
+
+    logging.basicConfig(level=level)
+    logging.getLogger("capa2yara").setLevel(level)
+
+    try:
+        rules = capa.main.get_rules(args.rules, disable_progress=True)
+        namespaces = capa.rules.index_rules_by_namespace(list(rules))
+        rules = capa.rules.RuleSet(rules)
+        logger.info("successfully loaded %s rules (including subscope rules which will be ignored)", len(rules))
+        if args.tag:
+            rules = rules.filter_rules_by_meta(args.tag)
+            logger.debug("selected %s rules", len(rules))
+            for i, r in enumerate(rules.rules, 1):
+                logger.debug(" %d. %s", i, r)
+    except (IOError, capa.rules.InvalidRule, capa.rules.InvalidRuleSet) as e:
+        logger.error("%s", str(e))
+        return -1
+
+    output_yar(
+        "// Rules from FireEye's https://github.com/fireeye/capa-rules converted to YARA using https://github.com/fireeye/capa/blob/master/scripts/capa2yara.py by Arnim Rupp"
+    )
+    output_yar(
+        "// Beware: These are less rules than capa (because not all fit into YARA, stats at EOF) and is less precise because e.g. capas function scopes are applied to the whole file"
+    )
+    output_yar(
+        '// Beware: Some rules are incomplete because an optional branch was not supported by YARA. These rules are marked in a comment in meta: (search for "incomplete")'
+    )
+    output_yar("// Rule authors and license stay the same")
+    output_yar(
+        '// att&ck and MBC tags are put into YARA rule tags. All rules are tagged with "CAPA" for easy filtering'
+    )
+    output_yar("// The date = in meta: is the date of converting (there is no date in capa rules)")
+    output_yar("// Minimum YARA version is 3.8.0 plus PE module")
+    output_yar('\nimport "pe"')
+
+    output_yar(condition_rule)
+
+    # do several rounds of converting rules because some rules for match: might not be converted in the 1st run
+    num_rules = 9999999
+    cround = 0
+    while num_rules != len(converted_rules) or cround < min_rounds:
+        cround += 1
+        logger.info("doing convert_rules(), round: " + str(cround))
+        num_rules = len(converted_rules)
+        convert_rules(rules, namespaces, cround)
+
+    # one last round to collect all unconverted rules
+    convert_rules(rules, namespaces, 9000)
+
+    stats = "\n// converted rules              : " + str(len(converted_rules))
+    stats += "\n//   among those are incomplete : " + str(count_incomplete)
+    stats += "\n// unconverted rules            : " + str(len(unsupported_capa_rules_list)) + "\n"
+    logger.info(stats)
+    output_yar(stats)
+
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

scripts/capa_as_library.py

@@ -6,18 +6,18 @@ import collections
 import capa.main
 import capa.rules
 import capa.engine
-import capa.render
 import capa.features
+import capa.render.json
 import capa.render.utils as rutils
+import capa.render.default
+import capa.render.result_document
 from capa.engine import *
-from capa.render import convert_capabilities_to_result_document
 
 # edit this to set the path for file to analyze and rule directory
 RULES_PATH = "/tmp/capa/rules/"
 
 # load rules from disk
-rules = capa.main.get_rules(RULES_PATH, disable_progress=True)
-rules = capa.rules.RuleSet(rules)
+rules = capa.rules.RuleSet(capa.main.get_rules(RULES_PATH, disable_progress=True))
 
 # == Render ddictionary helpers
 def render_meta(doc, ostream):
@@ -192,7 +192,7 @@ def render_dictionary(doc):
 def capa_details(file_path, output_format="dictionary"):
 
     # extract features and find capabilities
-    extractor = capa.main.get_extractor(file_path, "auto", capa.main.BACKEND_VIV, disable_progress=True)
+    extractor = capa.main.get_extractor(file_path, "auto", capa.main.BACKEND_VIV, [], False, disable_progress=True)
     capabilities, counts = capa.main.find_capabilities(rules, extractor, disable_progress=True)
 
     # collect metadata (used only to make rendering more complete)
@@ -202,14 +202,14 @@ def capa_details(file_path, output_format="dictionary"):
     capa_output = False
     if output_format == "dictionary":
         # ...as python dictionary, simplified as textable but in dictionary
-        doc = convert_capabilities_to_result_document(meta, rules, capabilities)
+        doc = capa.render.result_document.convert_capabilities_to_result_document(meta, rules, capabilities)
         capa_output = render_dictionary(doc)
     elif output_format == "json":
         # render results
         # ...as json
-        capa_output = json.loads(capa.render.render_json(meta, rules, capabilities))
+        capa_output = json.loads(capa.render.json.render(meta, rules, capabilities))
     elif output_format == "texttable":
         # ...as human readable text table
-        capa_output = capa.render.render_default(meta, rules, capabilities)
+        capa_output = capa.render.default.render(meta, rules, capabilities)
 
     return capa_output

scripts/capafmt.py

@@ -65,6 +65,8 @@ def main(argv=None):
             return 0
         else:
             logger.info("rule requires reformatting (%s)", rule.name)
+            if "\r\n" in rule.definition:
+                logger.info("please make sure that the file uses LF (\\n) line endings only")
             return 1
 
     if args.in_place:

scripts/ci.sh

@@ -9,6 +9,7 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 # Use a console with emojis support for a better experience
+# Use venv to ensure that `python` calls the correct python version
 
 # Stash uncommited changes
 MSG="pre-push-$(date +%s)";
@@ -25,17 +26,8 @@ restore_stashed() {
   fi
 }
 
-python_3() {
-  case "$(uname -s)" in
-   CYGWIN*|MINGW32*|MSYS*|MINGW*)
-     py -3 -m $1 > $2 2>&1;;
-   *)
-     python3 -m $1 > $2 2>&1;;
-  esac
-}
-
 # Run isort and print state
-python_3 'isort --profile black --length-sort --line-width 120 -c .' 'isort-output.log';
+python -m isort --profile black --length-sort --line-width 120 -c . > isort-output.log 2>&1;
 if [ $? == 0 ]; then
   echo 'isort succeeded!! 💖';
 else
@@ -46,7 +38,7 @@ else
 fi
 
 # Run black and print state
-python_3 'black -l 120 --check .' 'black-output.log';
+python -m black -l 120 --check . > black-output.log 2>&1;
 if [ $? == 0 ]; then
   echo 'black succeeded!! 💝';
 else
@@ -70,7 +62,7 @@ fi
 # Run tests except if first argument is no_tests
 if [ "$1" != 'no_tests' ]; then
   echo 'Running tests, please wait ⌛';
-  pytest tests/ --maxfail=1;
+  python -m pytest tests/ --maxfail=1;
   if [ $? == 0 ]; then
     echo 'Tests succeed!! 🎉';
   else

scripts/import-to-bn.py

@@ -25,7 +25,8 @@ Derived from: https://github.com/fireeye/capa/blob/master/scripts/import-to-ida.
 import os
 import json
 
-from binaryninja import *
+import binaryninja
+import binaryninja.interaction
 
 
 def append_func_cmt(bv, va, cmt):
@@ -46,31 +47,31 @@ def append_func_cmt(bv, va, cmt):
 def load_analysis(bv):
     shortname = os.path.splitext(os.path.basename(bv.file.filename))[0]
     dirname = os.path.dirname(bv.file.filename)
-    log_info(f"dirname: {dirname}\nshortname: {shortname}\n")
+    binaryninja.log_info(f"dirname: {dirname}\nshortname: {shortname}\n")
     if os.access(os.path.join(dirname, shortname + ".js"), os.R_OK):
         path = os.path.join(dirname, shortname + ".js")
     elif os.access(os.path.join(dirname, shortname + ".json"), os.R_OK):
         path = os.path.join(dirname, shortname + ".json")
     else:
-        path = interaction.get_open_filename_input("capa report:", "JSON (*.js *.json);;All Files (*)")
+        path = binaryninja.interaction.get_open_filename_input("capa report:", "JSON (*.js *.json);;All Files (*)")
     if not path or not os.access(path, os.R_OK):
-        log_error("Invalid filename.")
+        binaryninja.log_error("Invalid filename.")
         return 0
-    log_info("Using capa file %s" % path)
+    binaryninja.log_info("Using capa file %s" % path)
 
     with open(path, "rb") as f:
         doc = json.loads(f.read().decode("utf-8"))
 
     if "meta" not in doc or "rules" not in doc:
-        log_error("doesn't appear to be a capa report")
+        binaryninja.log_error("doesn't appear to be a capa report")
         return -1
 
     a = doc["meta"]["sample"]["md5"].lower()
-    md5 = Transform["MD5"]
-    rawhex = Transform["RawHex"]
+    md5 = binaryninja.Transform["MD5"]
+    rawhex = binaryninja.Transform["RawHex"]
     b = rawhex.encode(md5.encode(bv.parent_view.read(bv.parent_view.start, bv.parent_view.end))).decode("utf-8")
     if not a == b:
-        log_error("sample mismatch")
+        binaryninja.log_error("sample mismatch")
         return -2
 
     rows = []
@@ -96,7 +97,7 @@ def load_analysis(bv):
         else:
             cmt = "%s" % (name,)
 
-        log_info("0x%x: %s" % (va, cmt))
+        binaryninja.log_info("0x%x: %s" % (va, cmt))
         try:
             # message will look something like:
             #
@@ -105,7 +106,7 @@ def load_analysis(bv):
         except ValueError:
             continue
 
-    log_info("ok")
+    binaryninja.log_info("ok")
 
 
-PluginCommand.register("Load capa file", "Loads an analysis file from capa", load_analysis)
+binaryninja.PluginCommand.register("Load capa file", "Loads an analysis file from capa", load_analysis)

scripts/lint.py

@@ -25,18 +25,33 @@ import argparse
 import itertools
 import posixpath
 
+import termcolor
+import ruamel.yaml
+
 import capa.main
 import capa.rules
 import capa.engine
-import capa.features
 import capa.features.insn
+import capa.features.common
 
-logger = logging.getLogger("capa.lint")
+logger = logging.getLogger("lint")
 
 
-class Lint(object):
-    WARN = "WARN"
-    FAIL = "FAIL"
+def red(s):
+    return termcolor.colored(s, "red")
+
+
+def orange(s):
+    return termcolor.colored(s, "yellow")
+
+
+def green(s):
+    return termcolor.colored(s, "green")
+
+
+class Lint:
+    WARN = orange("WARN")
+    FAIL = red("FAIL")
 
     name = "lint"
     level = FAIL
@@ -179,6 +194,9 @@ class ExampleFileDNE(Lint):
         return not found
 
 
+DEFAULT_SIGNATURES = capa.main.get_default_signatures()
+
+
 class DoesntMatchExample(Lint):
     name = "doesn't match on referenced example"
     recommendation = "Fix the rule logic or provide a different example"
@@ -201,7 +219,9 @@ class DoesntMatchExample(Lint):
                 continue
 
             try:
-                extractor = capa.main.get_extractor(path, "auto", capa.main.BACKEND_VIV, disable_progress=True)
+                extractor = capa.main.get_extractor(
+                    path, "auto", capa.main.BACKEND_VIV, DEFAULT_SIGNATURES, False, disable_progress=True
+                )
                 capabilities, meta = capa.main.find_capabilities(ctx["rules"], extractor, disable_progress=True)
             except Exception as e:
                 logger.error("failed to extract capabilities: %s %s %s", rule.name, path, e)
@@ -211,6 +231,53 @@ class DoesntMatchExample(Lint):
                 return True
 
 
+class StatementWithSingleChildStatement(Lint):
+    name = "rule contains one or more statements with a single child statement"
+    recommendation = "remove the superfluous parent statement"
+    recommendation_template = "remove the superfluous parent statement: {:s}"
+    violation = False
+
+    def check_rule(self, ctx, rule):
+        self.violation = False
+
+        def rec(statement, is_root=False):
+            if isinstance(statement, (capa.engine.And, capa.engine.Or)):
+                children = list(statement.get_children())
+                if not is_root and len(children) == 1 and isinstance(children[0], capa.engine.Statement):
+                    self.recommendation = self.recommendation_template.format(str(statement))
+                    self.violation = True
+                for child in children:
+                    rec(child)
+
+        rec(rule.statement, is_root=True)
+
+        return self.violation
+
+
+class OrStatementWithAlwaysTrueChild(Lint):
+    name = "rule contains an `or` statement that's always True because of an `optional` or other child statement that's always True"
+    recommendation = "clarify the rule logic, e.g. by moving the always True child statement"
+    recommendation_template = "clarify the rule logic, e.g. by moving the always True child statement: {:s}"
+    violation = False
+
+    def check_rule(self, ctx, rule):
+        self.violation = False
+
+        def rec(statement):
+            if isinstance(statement, capa.engine.Or):
+                children = list(statement.get_children())
+                for child in children:
+                    # `Some` implements `optional` which is an alias for `0 or more`
+                    if isinstance(child, capa.engine.Some) and child.count == 0:
+                        self.recommendation = self.recommendation_template.format(str(child))
+                        self.violation = True
+                    rec(child)
+
+        rec(rule.statement)
+
+        return self.violation
+
+
 class UnusualMetaField(Lint):
     name = "unusual meta field"
     recommendation = "Remove the meta field"
@@ -259,7 +326,7 @@ class FeatureStringTooShort(Lint):
 
     def check_features(self, ctx, features):
         for feature in features:
-            if isinstance(feature, capa.features.String):
+            if isinstance(feature, capa.features.common.String):
                 if len(feature.value) < 4:
                     self.recommendation = self.recommendation.format(feature.value)
                     return True
@@ -301,6 +368,16 @@ class FeatureNtdllNtoskrnlApi(Lint):
         return False
 
 
+class FormatLineFeedEOL(Lint):
+    name = "line(s) end with CRLF (\\r\\n)"
+    recommendation = "convert line endings to LF (\\n) for example using dos2unix"
+
+    def check_rule(self, ctx, rule):
+        if len(rule.definition.split("\r\n")) > 0:
+            return False
+        return True
+
+
 class FormatSingleEmptyLineEOF(Lint):
     name = "EOF format"
     recommendation = "end file with a single empty line"
@@ -321,12 +398,43 @@ class FormatIncorrect(Lint):
 
         if actual != expected:
             diff = difflib.ndiff(actual.splitlines(1), expected.splitlines(True))
-            self.recommendation = self.recommendation_template.format("".join(diff))
+            recommendation_template = self.recommendation_template
+            if "\r\n" in actual:
+                recommendation_template = (
+                    self.recommendation_template + "\nplease make sure that the file uses LF (\\n) line endings only"
+                )
+            self.recommendation = recommendation_template.format("".join(diff))
             return True
 
         return False
 
 
+class FormatStringQuotesIncorrect(Lint):
+    name = "rule string quotes incorrect"
+
+    def check_rule(self, ctx, rule):
+        events = capa.rules.Rule._get_ruamel_yaml_parser().parse(rule.definition)
+        for key in events:
+            if not (isinstance(key, ruamel.yaml.ScalarEvent) and key.value == "string"):
+                continue
+            value = next(events)  # assume value is next event
+            if not isinstance(value, ruamel.yaml.ScalarEvent):
+                # ignore non-scalar
+                continue
+            if value.value.startswith("/") and value.value.endswith(("/", "/i")):
+                # ignore regex for now
+                continue
+            if value.style is None:
+                # no quotes
+                self.recommendation = 'add double quotes to "%s"' % value.value
+                return True
+            if value.style == "'":
+                # single quote
+                self.recommendation = 'change single quotes to double quotes for "%s"' % value.value
+                return True
+        return False
+
+
 def run_lints(lints, ctx, rule):
     for lint in lints:
         if lint.check_rule(ctx, rule):
@@ -385,7 +493,9 @@ def lint_features(ctx, rule):
 
 
 FORMAT_LINTS = (
+    FormatLineFeedEOL(),
     FormatSingleEmptyLineEOF(),
+    FormatStringQuotesIncorrect(),
     FormatIncorrect(),
 )
 
@@ -401,7 +511,7 @@ def get_normpath(path):
 def get_features(ctx, rule):
     # get features from rule and all dependencies including subscopes and matched rules
     features = []
-    namespaces = capa.rules.index_rules_by_namespace([rule])
+    namespaces = ctx["rules"].rules_by_namespace
     deps = [ctx["rules"].rules[dep] for dep in rule.get_dependencies(namespaces)]
     for r in [rule] + deps:
         features.extend(get_rule_features(r))
@@ -422,7 +532,11 @@ def get_rule_features(rule):
     return features
 
 
-LOGIC_LINTS = (DoesntMatchExample(),)
+LOGIC_LINTS = (
+    DoesntMatchExample(),
+    StatementWithSingleChildStatement(),
+    OrStatementWithAlwaysTrueChild(),
+)
 
 
 def lint_logic(ctx, rule):
@@ -478,15 +592,37 @@ def lint_rule(ctx, rule):
 
         print("")
 
-    lints_failed = any(map(lambda v: v.level == Lint.FAIL, violations))
+    if is_nursery_rule(rule):
+        has_examples = not any(map(lambda v: v.level == Lint.FAIL and v.name == "missing examples", violations))
+        lints_failed = len(
+            tuple(
+                filter(
+                    lambda v: v.level == Lint.FAIL
+                    and not (v.name == "missing examples" or v.name == "referenced example doesn't exist"),
+                    violations,
+                )
+            )
+        )
+        lints_warned = len(
+            tuple(
+                filter(
+                    lambda v: v.level == Lint.WARN
+                    or (v.level == Lint.FAIL and v.name == "referenced example doesn't exist"),
+                    violations,
+                )
+            )
+        )
 
-    if not lints_failed and is_nursery_rule(rule):
-        print("")
-        print("%s%s" % ("    (nursery) ", rule.name))
-        print("%s  %s: %s: %s" % ("    ", Lint.WARN, "no lint failures", "Graduate the rule"))
-        print("")
+        if (not lints_failed) and (not lints_warned) and has_examples:
+            print("")
+            print("%s%s" % ("    (nursery) ", rule.name))
+            print("%s  %s: %s: %s" % ("    ", Lint.WARN, green("no lint failures"), "Graduate the rule"))
+            print("")
+    else:
+        lints_failed = len(tuple(filter(lambda v: v.level == Lint.FAIL, violations)))
+        lints_warned = len(tuple(filter(lambda v: v.level == Lint.WARN, violations)))
 
-    return lints_failed and not is_nursery_rule(rule)
+    return (lints_failed, lints_warned)
 
 
 def lint(ctx, rules):
@@ -496,15 +632,20 @@ def lint(ctx, rules):
         for each sample, record sample id of sha256, md5, and filename.
         see `collect_samples(path)`.
       rules (List[Rule]): the rules to lint.
+
+    Returns: Dict[string, Tuple(int, int)]
+      - # lints failed
+      - # lints warned
     """
-    did_suggest_fix = False
-    for rule in rules.rules.values():
+    ret = {}
+
+    for name, rule in rules.rules.items():
         if rule.meta.get("capa/subscope-rule", False):
             continue
 
-        did_suggest_fix = lint_rule(ctx, rule) or did_suggest_fix
+        ret[name] = lint_rule(ctx, rule)
 
-    return did_suggest_fix
+    return ret
 
 
 def collect_samples(path):
@@ -554,7 +695,7 @@ def main(argv=None):
 
     samples_path = os.path.join(os.path.dirname(__file__), "..", "tests", "data")
 
-    parser = argparse.ArgumentParser(description="A program.")
+    parser = argparse.ArgumentParser(description="Lint capa rules.")
     capa.main.install_common_args(parser, wanted={"tag"})
     parser.add_argument("rules", type=str, help="Path to rules")
     parser.add_argument("--samples", type=str, default=samples_path, help="Path to samples")
@@ -566,8 +707,12 @@ def main(argv=None):
     args = parser.parse_args(args=argv)
     capa.main.handle_common_args(args)
 
-    logging.getLogger("capa").setLevel(logging.CRITICAL)
-    logging.getLogger("viv_utils").setLevel(logging.CRITICAL)
+    if args.debug:
+        logging.getLogger("capa").setLevel(logging.DEBUG)
+        logging.getLogger("viv_utils").setLevel(logging.DEBUG)
+    else:
+        logging.getLogger("capa").setLevel(logging.ERROR)
+        logging.getLogger("viv_utils").setLevel(logging.ERROR)
 
     time0 = time.time()
 
@@ -597,16 +742,33 @@ def main(argv=None):
         "is_thorough": args.thorough,
     }
 
-    did_violate = lint(ctx, rules)
+    results_by_name = lint(ctx, rules)
+    failed_rules = []
+    warned_rules = []
+    for name, (fail_count, warn_count) in results_by_name.items():
+        if fail_count > 0:
+            failed_rules.append(name)
+
+        if warn_count > 0:
+            warned_rules.append(name)
 
     min, sec = divmod(time.time() - time0, 60)
     logger.debug("lints ran for ~ %02d:%02dm", min, sec)
 
-    if not did_violate:
-        logger.info("no lints failed, nice!")
-        return 0
-    else:
+    if warned_rules:
+        print(orange("rules with WARN:"))
+        for warned_rule in sorted(warned_rules):
+            print("  - " + warned_rule)
+        print()
+
+    if failed_rules:
+        print(red("rules with FAIL:"))
+        for failed_rule in sorted(failed_rules):
+            print("  - " + failed_rule)
         return 1
+    else:
+        logger.info(green("no lints failed, nice!"))
+        return 0
 
 
 if __name__ == "__main__":

scripts/match-function-id.py

@@ -0,0 +1,134 @@
+#!/usr/bin/env python3
+"""
+Copyright (C) 2021 FireEye, Inc. All Rights Reserved.
+Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+You may obtain a copy of the License at: [package root]/LICENSE.txt
+Unless required by applicable law or agreed to in writing, software distributed under the License
+ is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and limitations under the License.
+
+match-function-id
+
+Show the names of functions as recognized by the function identification subsystem.
+This can help identify library functions statically linked into a program,
+such as when triaging false positive matches in capa rules.
+
+Example::
+
+    $ python scripts/match-function-id.py --signature sigs/vc6.pat.gz /tmp/suspicious.dll_
+    0x44cf30: ?GetPdbDll@@YAPAUHINSTANCE__@@XZ
+    0x44bb20: ?_strlen_priv@@YAIPBD@Z
+    0x44b6b0: ?invoke_main@@YAHXZ
+    0x44a5d0: ?find_pe_section@@YAPAU_IMAGE_SECTION_HEADER@@QAEI@Z
+    0x44a690: ?is_potentially_valid_image_base@@YA_NQAX@Z
+    0x44cbe0: ___get_entropy
+    0x44a4a0: __except_handler4
+    0x44b3d0: ?pre_cpp_initialization@@YAXXZ
+    0x44b2e0: ?pre_c_initialization@@YAHXZ
+    0x44b3c0: ?post_pgo_initialization@@YAHXZ
+    0x420156: ?
+    0x420270: ?
+    0x430dcd: ?
+    0x44d930: __except_handler4_noexcept
+    0x41e960: ?
+    0x44a1e0: @_RTC_AllocaHelper@12
+    0x44ba90: ?_getMemBlockDataString@@YAXPAD0PBDI@Z
+    0x44a220: @_RTC_CheckStackVars2@12
+    0x44a790: ___scrt_dllmain_after_initialize_c
+    0x44a7d0: ___scrt_dllmain_before_initialize_c
+    0x44a800: ___scrt_dllmain_crt_thread_attach
+    0x44a860: ___scrt_dllmain_exception_filter
+    0x44a900: ___scrt_dllmain_uninitialize_critical
+    0x44ad10: _at_quick_exit
+    0x44b940: ?_RTC_Failure@@YAXPAXH@Z
+    0x44be60: __RTC_UninitUse
+    0x44bfd0: __RTC_GetErrDesc
+    0x44c060: __RTC_SetErrorType
+    0x44cb60: ?
+    0x44cba0: __guard_icall_checks_enforced
+"""
+import sys
+import logging
+import argparse
+
+import flirt
+import viv_utils
+import viv_utils.flirt
+
+import capa.main
+import capa.rules
+import capa.engine
+import capa.helpers
+import capa.features
+import capa.features.freeze
+
+logger = logging.getLogger("capa.match-function-id")
+
+
+def main(argv=None):
+    if argv is None:
+        argv = sys.argv[1:]
+
+    parser = argparse.ArgumentParser(description="FLIRT match each function")
+    parser.add_argument("sample", type=str, help="Path to sample to analyze")
+    parser.add_argument(
+        "-F",
+        "--function",
+        type=lambda x: int(x, 0x10),
+        help="match a specific function by VA, rather than add functions",
+    )
+    parser.add_argument(
+        "--signature",
+        action="append",
+        dest="signatures",
+        type=str,
+        default=[],
+        help="use the given signatures to identify library functions, file system paths to .sig/.pat files.",
+    )
+    parser.add_argument("-d", "--debug", action="store_true", help="Enable debugging output on STDERR")
+    parser.add_argument("-q", "--quiet", action="store_true", help="Disable all output but errors")
+    args = parser.parse_args(args=argv)
+
+    if args.quiet:
+        logging.basicConfig(level=logging.ERROR)
+        logging.getLogger().setLevel(logging.ERROR)
+    elif args.debug:
+        logging.basicConfig(level=logging.DEBUG)
+        logging.getLogger().setLevel(logging.DEBUG)
+    else:
+        logging.basicConfig(level=logging.INFO)
+        logging.getLogger().setLevel(logging.INFO)
+
+    # disable vivisect-related logging, it's verbose and not relevant for capa users
+    capa.main.set_vivisect_log_level(logging.CRITICAL)
+
+    analyzers = []
+    for sigpath in args.signatures:
+        sigs = capa.main.load_flirt_signature(sigpath)
+
+        with capa.main.timing("flirt: compiling sigs"):
+            matcher = flirt.compile(sigs)
+
+        analyzer = viv_utils.flirt.FlirtFunctionAnalyzer(matcher, sigpath)
+        logger.debug("registering viv function analyzer: %s", repr(analyzer))
+        analyzers.append(analyzer)
+
+    vw = viv_utils.getWorkspace(args.sample, analyze=True, should_save=False)
+
+    functions = vw.getFunctions()
+    if args.function:
+        functions = [args.function]
+
+    for function in functions:
+        logger.debug("matching function: 0x%04x", function)
+        for analyzer in analyzers:
+            name = viv_utils.flirt.match_function_flirt_signatures(analyzer.matcher, vw, function)
+            if name:
+                print("0x%04x: %s" % (function, name))
+
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

scripts/show-capabilities-by-function.py

@@ -59,10 +59,10 @@ import colorama
 import capa.main
 import capa.rules
 import capa.engine
-import capa.render
 import capa.features
 import capa.render.utils as rutils
 import capa.features.freeze
+import capa.render.result_document
 from capa.helpers import get_file_taste
 
 logger = logging.getLogger("capa.show-capabilities-by-function")
@@ -111,7 +111,7 @@ def main(argv=None):
         argv = sys.argv[1:]
 
     parser = argparse.ArgumentParser(description="detect capabilities in programs.")
-    capa.main.install_common_args(parser, wanted={"format", "sample", "rules", "tag"})
+    capa.main.install_common_args(parser, wanted={"format", "backend", "sample", "signatures", "rules", "tag"})
     args = parser.parse_args(args=argv)
     capa.main.handle_common_args(args)
 
@@ -121,20 +121,6 @@ def main(argv=None):
         logger.error("%s", str(e))
         return -1
 
-    if args.rules == "(embedded rules)":
-        logger.info("-" * 80)
-        logger.info(" Using default embedded rules.")
-        logger.info(" To provide your own rules, use the form `capa.exe -r ./path/to/rules/  /path/to/mal.exe`.")
-        logger.info(" You can see the current default rule set here:")
-        logger.info("     https://github.com/fireeye/capa-rules")
-        logger.info("-" * 80)
-
-        logger.debug("detected running from source")
-        args.rules = os.path.join(os.path.dirname(__file__), "..", "rules")
-        logger.debug("default rule path (source method): %s", args.rules)
-    else:
-        logger.info("using rules path: %s", args.rules)
-
     try:
         rules = capa.main.get_rules(args.rules)
         rules = capa.rules.RuleSet(rules)
@@ -146,14 +132,24 @@ def main(argv=None):
         logger.error("%s", str(e))
         return -1
 
+    try:
+        sig_paths = capa.main.get_signatures(args.signatures)
+    except (IOError) as e:
+        logger.error("%s", str(e))
+        return -1
+
     if (args.format == "freeze") or (args.format == "auto" and capa.features.freeze.is_freeze(taste)):
         format = "freeze"
         with open(args.sample, "rb") as f:
             extractor = capa.features.freeze.load(f.read())
     else:
         format = args.format
+        should_save_workspace = os.environ.get("CAPA_SAVE_WORKSPACE") not in ("0", "no", "NO", "n", None)
+
         try:
-            extractor = capa.main.get_extractor(args.sample, args.format)
+            extractor = capa.main.get_extractor(
+                args.sample, args.format, args.backend, sig_paths, should_save_workspace
+            )
         except capa.main.UnsupportedFormatError:
             logger.error("-" * 80)
             logger.error(" Input file does not appear to be a PE file.")
@@ -190,7 +186,7 @@ def main(argv=None):
     #  - when not an interactive session, and disable coloring
     # renderers should use coloring and assume it will be stripped out if necessary.
     colorama.init()
-    doc = capa.render.convert_capabilities_to_result_document(meta, rules, capabilities)
+    doc = capa.render.result_document.convert_capabilities_to_result_document(meta, rules, capabilities)
     print(render_matches_by_function(doc))
     colorama.deinit()
 

scripts/show-features.py

@@ -64,8 +64,10 @@ Example::
     insn: 0x10001027: mnemonic(shl)
     ...
 """
+import os
 import sys
 import logging
+import os.path
 import argparse
 
 import capa.main
@@ -83,7 +85,7 @@ def main(argv=None):
         argv = sys.argv[1:]
 
     parser = argparse.ArgumentParser(description="Show the features that capa extracts from the given sample")
-    capa.main.install_common_args(parser, wanted={"format", "sample"})
+    capa.main.install_common_args(parser, wanted={"format", "sample", "signatures"})
 
     parser.add_argument("-F", "--function", type=lambda x: int(x, 0x10), help="Show features for specific function")
     args = parser.parse_args(args=argv)
@@ -95,12 +97,21 @@ def main(argv=None):
         logger.error("%s", str(e))
         return -1
 
+    try:
+        sig_paths = capa.main.get_signatures(args.signatures)
+    except (IOError) as e:
+        logger.error("%s", str(e))
+        return -1
+
     if (args.format == "freeze") or (args.format == "auto" and capa.features.freeze.is_freeze(taste)):
         with open(args.sample, "rb") as f:
             extractor = capa.features.freeze.load(f.read())
     else:
+        should_save_workspace = os.environ.get("CAPA_SAVE_WORKSPACE") not in ("0", "no", "NO", "n", None)
         try:
-            extractor = capa.main.get_extractor(args.sample, args.format, capa.main.BACKEND_VIV)
+            extractor = capa.main.get_extractor(
+                args.sample, args.format, capa.main.BACKEND_VIV, sig_paths, should_save_workspace
+            )
         except capa.main.UnsupportedFormatError:
             logger.error("-" * 80)
             logger.error(" Input file does not appear to be a PE file.")
@@ -135,9 +146,9 @@ def main(argv=None):
         if args.format == "freeze":
             functions = tuple(filter(lambda f: f == args.function, functions))
         else:
-            functions = tuple(filter(lambda f: capa.helpers.oint(f) == args.function, functions))
+            functions = tuple(filter(lambda f: int(f) == args.function, functions))
 
-            if args.function not in [capa.helpers.oint(f) for f in functions]:
+            if args.function not in [int(f) for f in functions]:
                 print("0x%X not a function" % args.function)
                 return -1
 
@@ -151,10 +162,14 @@ def main(argv=None):
 
 
 def ida_main():
+    import idc
+
+    import capa.features.extractors.ida.extractor
+
     function = idc.get_func_attr(idc.here(), idc.FUNCATTR_START)
     print("getting features for current function 0x%X" % function)
 
-    extractor = capa.features.extractors.ida.IdaFeatureExtractor()
+    extractor = capa.features.extractors.ida.extractor.IdaFeatureExtractor()
 
     if not function:
         for feature, va in extractor.extract_file_features():
@@ -180,6 +195,13 @@ def ida_main():
 
 def print_features(functions, extractor):
     for f in functions:
+        function_address = int(f)
+
+        if extractor.is_library_function(function_address):
+            function_name = extractor.get_function_name(function_address)
+            logger.debug("skipping library function 0x%x (%s)", function_address, function_name)
+            continue
+
         for feature, va in extractor.extract_function_features(f):
             print("func: 0x%08x: %s" % (va, feature))
 

setup.py

@@ -7,38 +7,27 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 import os
-import sys
 
 import setuptools
 
 requirements = [
-    "six",
-    "tqdm",
-    "pyyaml",
-    "tabulate",
-    "colorama",
-    "termcolor",
-    "ruamel.yaml",
-    "wcwidth",
+    "tqdm==4.61.2",
+    "pyyaml==5.4.1",
+    "tabulate==0.8.9",
+    "colorama==0.4.4",
+    "termcolor==1.1.0",
+    "wcwidth==0.2.5",
     "ida-settings==2.1.0",
+    "viv-utils[flirt]==0.6.5",
+    "halo==0.0.31",
+    "networkx==2.5.1",
+    "ruamel.yaml==0.17.10",
+    "vivisect==1.0.3",
+    "smda==1.5.18",
+    "pefile==2021.5.24",
+    "typing==3.7.4.3",
 ]
 
-if sys.version_info >= (3, 0):
-    # py3
-    requirements.append("halo")
-    requirements.append("networkx")
-    requirements.append("vivisect==1.0.0")
-    requirements.append("viv-utils==0.3.19")
-    requirements.append("smda==1.5.13")
-else:
-    # py2
-    requirements.append("enum34==1.1.6")  # v1.1.6 is needed by halo 0.0.30 / spinners 0.0.24
-    requirements.append("halo==0.0.30")  # halo==0.0.30 is the last version to support py2.7
-    requirements.append("vivisect==0.2.1")
-    requirements.append("viv-utils==0.3.19")
-    requirements.append("networkx==2.2")  # v2.2 is last version supported by Python 2.7
-    requirements.append("backports.functools-lru-cache")
-
 # this sets __version__
 # via: http://stackoverflow.com/a/7071358/87207
 # and: http://stackoverflow.com/a/2073599/87207
@@ -77,14 +66,21 @@ setuptools.setup(
     install_requires=requirements,
     extras_require={
         "dev": [
-            "pytest",
-            "pytest-sugar",
-            "pytest-instafail",
-            "pytest-cov",
-            "pycodestyle",
-            "black ; python_version>'3.0'",
-            "isort",
-        ]
+            "pytest==6.2.4",
+            "pytest-sugar==0.9.4",
+            "pytest-instafail==0.4.2",
+            "pytest-cov==2.12.1",
+            "pycodestyle==2.7.0",
+            "black==21.7b0",
+            "isort==5.9.2",
+            "mypy==0.910",
+            # type stubs for mypy
+            "types-backports==0.1.3",
+            "types-colorama==0.4.2",
+            "types-PyYAML==5.4.3",
+            "types-tabulate==0.1.1",
+            "types-termcolor==0.1.1",
+        ],
     },
     zip_safe=False,
     keywords="capa malware analysis capability detection FLARE",
@@ -94,8 +90,8 @@ setuptools.setup(
         "Intended Audience :: Information Technology",
         "License :: OSI Approved :: Apache Software License",
         "Natural Language :: English",
-        "Programming Language :: Python :: 2.7",
         "Programming Language :: Python :: 3",
         "Topic :: Security",
     ],
+    python_requires=">=3.6",
 )

sigs/README.md

@@ -0,0 +1,9 @@
+# capa/sigs
+
+This directory contains FLIRT signatures that capa uses to identify library functions.
+Typically, capa will ignore library functions, which reduces false positives and improves runtime.
+
+These FLIRT signatures were generated by FireEye using the Hex-Rays FLAIR tools such as `pcf` and `sigmake`.
+FireEye generated the signatures from source data that they collected; these signatures are not derived from the FLIRT signatures distributed with IDA Pro.
+
+The signatures in this directory have the same license as capa: Apache 2.0.

tests/fixtures.py

@@ -8,25 +8,20 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 import os
-import sys
 import os.path
 import binascii
 import contextlib
 import collections
+from functools import lru_cache
 
 import pytest
 
 import capa.main
 import capa.features.file
 import capa.features.insn
+import capa.features.common
 import capa.features.basicblock
-from capa.features import ARCH_X32, ARCH_X64
-
-try:
-    from functools import lru_cache
-except ImportError:
-    from backports.functools_lru_cache import lru_cache
-
+from capa.features.common import ARCH_X32, ARCH_X64
 
 CD = os.path.dirname(__file__)
 
@@ -41,11 +36,11 @@ def xfail(condition, reason=None):
     example::
 
         # this test:
-        #  - passes on py3 if foo() works
-        #  - fails  on py3 if foo() fails
-        #  - xfails on py2 if foo() fails
-        #  - fails  on py2 if foo() works
-        with xfail(sys.version_info < (3, 0), reason="py2 doesn't foo"):
+        #  - passes on Linux if foo() works
+        #  - fails  on Linux if foo() fails
+        #  - xfails on Windows if foo() fails
+        #  - fails  on Windows if foo() works
+        with xfail(sys.platform == "win32", reason="doesn't work on Windows"):
             foo()
     """
     try:
@@ -69,17 +64,28 @@ def xfail(condition, reason=None):
             raise RuntimeError("expected to fail, but didn't")
 
 
-@lru_cache()
+# need to limit cache size so GitHub Actions doesn't run out of memory, see #545
+@lru_cache(maxsize=1)
 def get_viv_extractor(path):
-    import capa.features.extractors.viv
+    import capa.main
+    import capa.features.extractors.viv.extractor
+
+    sigpaths = [
+        os.path.join(CD, "data", "sigs", "test_aulldiv.pat"),
+        os.path.join(CD, "data", "sigs", "test_aullrem.pat.gz"),
+        os.path.join(CD, "..", "sigs", "1_flare_msvc_rtf_32_64.sig"),
+        os.path.join(CD, "..", "sigs", "2_flare_msvc_atlmfc_32_64.sig"),
+        os.path.join(CD, "..", "sigs", "3_flare_common_libs.sig"),
+    ]
 
     if "raw32" in path:
-        vw = capa.main.get_workspace(path, "sc32", should_save=False)
+        vw = capa.main.get_workspace(path, "sc32", sigpaths=sigpaths)
     elif "raw64" in path:
-        vw = capa.main.get_workspace(path, "sc64", should_save=False)
+        vw = capa.main.get_workspace(path, "sc64", sigpaths=sigpaths)
     else:
-        vw = capa.main.get_workspace(path, "auto", should_save=True)
-    extractor = capa.features.extractors.viv.VivisectFeatureExtractor(vw, path)
+        vw = capa.main.get_workspace(path, "auto", sigpaths=sigpaths)
+    vw.saveWorkspace()
+    extractor = capa.features.extractors.viv.extractor.VivisectFeatureExtractor(vw, path)
     fixup_viv(path, extractor)
     return extractor
 
@@ -98,14 +104,21 @@ def get_smda_extractor(path):
     from smda.SmdaConfig import SmdaConfig
     from smda.Disassembler import Disassembler
 
-    import capa.features.extractors.smda
+    import capa.features.extractors.smda.extractor
 
     config = SmdaConfig()
     config.STORE_BUFFER = True
     disasm = Disassembler(config)
     report = disasm.disassembleFile(path)
 
-    return capa.features.extractors.smda.SmdaFeatureExtractor(report, path)
+    return capa.features.extractors.smda.extractor.SmdaFeatureExtractor(report, path)
+
+
+@lru_cache(maxsize=1)
+def get_pefile_extractor(path):
+    import capa.features.extractors.pefile
+
+    return capa.features.extractors.pefile.PefileFeatureExtractor(path)
 
 
 @lru_cache()
@@ -141,6 +154,7 @@ def extract_basic_block_features(extractor, f, bb):
     return features
 
 
+# note: too reduce the testing time it's recommended to reuse already existing test samples, if possible
 def get_data_path_by_name(name):
     if name == "mimikatz":
         return os.path.join(CD, "data", "mimikatz.exe_")
@@ -242,14 +256,14 @@ def sample(request):
 
 def get_function(extractor, fva):
     for f in extractor.get_functions():
-        if f.__int__() == fva:
+        if int(f) == fva:
             return f
     raise ValueError("function not found")
 
 
 def get_basic_block(extractor, f, va):
     for bb in extractor.get_basic_blocks(f):
-        if bb.__int__() == va:
+        if int(bb) == va:
             return bb
     raise ValueError("basic block not found")
 
@@ -257,34 +271,34 @@ def get_basic_block(extractor, f, va):
 def resolve_scope(scope):
     if scope == "file":
 
-        def inner(extractor):
+        def inner_file(extractor):
             return extract_file_features(extractor)
 
-        inner.__name__ = scope
-        return inner
+        inner_file.__name__ = scope
+        return inner_file
     elif "bb=" in scope:
         # like `function=0x401000,bb=0x40100A`
         fspec, _, bbspec = scope.partition(",")
         fva = int(fspec.partition("=")[2], 0x10)
         bbva = int(bbspec.partition("=")[2], 0x10)
 
-        def inner(extractor):
+        def inner_bb(extractor):
             f = get_function(extractor, fva)
             bb = get_basic_block(extractor, f, bbva)
             return extract_basic_block_features(extractor, f, bb)
 
-        inner.__name__ = scope
-        return inner
+        inner_bb.__name__ = scope
+        return inner_bb
     elif scope.startswith("function"):
         # like `function=0x401000`
         va = int(scope.partition("=")[2], 0x10)
 
-        def inner(extractor):
+        def inner_function(extractor):
             f = get_function(extractor, va)
             return extract_function_features(extractor, f)
 
-        inner.__name__ = scope
-        return inner
+        inner_function.__name__ = scope
+        return inner_function
     else:
         raise ValueError("unexpected scope fixture")
 
@@ -310,177 +324,186 @@ def parametrize(params, values, **kwargs):
     return pytest.mark.parametrize(params, values, ids=ids, **kwargs)
 
 
-FEATURE_PRESENCE_TESTS = [
-    # file/characteristic("embedded pe")
-    ("pma12-04", "file", capa.features.Characteristic("embedded pe"), True),
-    # file/string
-    ("mimikatz", "file", capa.features.String("SCardControl"), True),
-    ("mimikatz", "file", capa.features.String("SCardTransmit"), True),
-    ("mimikatz", "file", capa.features.String("ACR  > "), True),
-    ("mimikatz", "file", capa.features.String("nope"), False),
-    # file/sections
-    ("mimikatz", "file", capa.features.file.Section(".text"), True),
-    ("mimikatz", "file", capa.features.file.Section(".nope"), False),
-    # IDA doesn't extract unmapped sections by default
-    # ("mimikatz", "file", capa.features.file.Section(".rsrc"), True),
-    # file/exports
-    ("kernel32", "file", capa.features.file.Export("BaseThreadInitThunk"), True),
-    ("kernel32", "file", capa.features.file.Export("lstrlenW"), True),
-    ("kernel32", "file", capa.features.file.Export("nope"), False),
-    # file/imports
-    ("mimikatz", "file", capa.features.file.Import("advapi32.CryptSetHashParam"), True),
-    ("mimikatz", "file", capa.features.file.Import("CryptSetHashParam"), True),
-    ("mimikatz", "file", capa.features.file.Import("kernel32.IsWow64Process"), True),
-    ("mimikatz", "file", capa.features.file.Import("msvcrt.exit"), True),
-    ("mimikatz", "file", capa.features.file.Import("cabinet.#11"), True),
-    ("mimikatz", "file", capa.features.file.Import("#11"), False),
-    ("mimikatz", "file", capa.features.file.Import("#nope"), False),
-    ("mimikatz", "file", capa.features.file.Import("nope"), False),
-    ("mimikatz", "file", capa.features.file.Import("advapi32.CryptAcquireContextW"), True),
-    ("mimikatz", "file", capa.features.file.Import("advapi32.CryptAcquireContext"), True),
-    ("mimikatz", "file", capa.features.file.Import("CryptAcquireContextW"), True),
-    ("mimikatz", "file", capa.features.file.Import("CryptAcquireContext"), True),
-    # function/characteristic(loop)
-    ("mimikatz", "function=0x401517", capa.features.Characteristic("loop"), True),
-    ("mimikatz", "function=0x401000", capa.features.Characteristic("loop"), False),
-    # bb/characteristic(tight loop)
-    ("mimikatz", "function=0x402EC4", capa.features.Characteristic("tight loop"), True),
-    ("mimikatz", "function=0x401000", capa.features.Characteristic("tight loop"), False),
-    # bb/characteristic(stack string)
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("stack string"), True),
-    ("mimikatz", "function=0x401000", capa.features.Characteristic("stack string"), False),
-    # bb/characteristic(tight loop)
-    ("mimikatz", "function=0x402EC4,bb=0x402F8E", capa.features.Characteristic("tight loop"), True),
-    ("mimikatz", "function=0x401000,bb=0x401000", capa.features.Characteristic("tight loop"), False),
-    # insn/mnemonic
-    ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("push"), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("movzx"), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("xor"), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("in"), False),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("out"), False),
-    # insn/number
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0x3136B0), True),
-    # insn/number: stack adjustments
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xC), False),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0x10), False),
-    # insn/number: arch flavors
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF, arch=ARCH_X32), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF, arch=ARCH_X64), False),
-    # insn/offset
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x4), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0xC), True),
-    # insn/offset, issue #276
-    ("64d9f", "function=0x10001510,bb=0x100015B0", capa.features.insn.Offset(0x4000), True),
-    # insn/offset: stack references
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x8), False),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x10), False),
-    # insn/offset: negative
-    ("mimikatz", "function=0x4011FB", capa.features.insn.Offset(-0x1), True),
-    ("mimikatz", "function=0x4011FB", capa.features.insn.Offset(-0x2), True),
-    # insn/offset: arch flavors
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0, arch=ARCH_X32), True),
-    ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0, arch=ARCH_X64), False),
-    # insn/api
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptAcquireContextW"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptAcquireContext"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptGenKey"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptImportKey"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptDestroyKey"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptAcquireContextW"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptAcquireContext"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptGenKey"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptImportKey"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptDestroyKey"), True),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("Nope"), False),
-    ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.Nope"), False),
-    # insn/api: thunk
-    ("mimikatz", "function=0x4556E5", capa.features.insn.API("advapi32.LsaQueryInformationPolicy"), True),
-    ("mimikatz", "function=0x4556E5", capa.features.insn.API("LsaQueryInformationPolicy"), True),
-    # insn/api: x64
-    (
-        "kernel32-64",
-        "function=0x180001010",
-        capa.features.insn.API("RtlVirtualUnwind"),
-        True,
-    ),
-    ("kernel32-64", "function=0x180001010", capa.features.insn.API("RtlVirtualUnwind"), True),
-    # insn/api: x64 thunk
-    (
-        "kernel32-64",
-        "function=0x1800202B0",
-        capa.features.insn.API("RtlCaptureContext"),
-        True,
-    ),
-    ("kernel32-64", "function=0x1800202B0", capa.features.insn.API("RtlCaptureContext"), True),
-    # insn/api: x64 nested thunk
-    ("al-khaser x64", "function=0x14004B4F0", capa.features.insn.API("__vcrt_GetModuleHandle"), True),
-    # insn/api: call via jmp
-    ("mimikatz", "function=0x40B3C6", capa.features.insn.API("LocalFree"), True),
-    ("c91887...", "function=0x40156F", capa.features.insn.API("CloseClipboard"), True),
-    # TODO ignore thunk functions that call via jmp?
-    # insn/api: resolve indirect calls
-    ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.CreatePipe"), True),
-    ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.SetHandleInformation"), True),
-    ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.CloseHandle"), True),
-    ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.WriteFile"), True),
-    # insn/string
-    ("mimikatz", "function=0x40105D", capa.features.String("SCardControl"), True),
-    ("mimikatz", "function=0x40105D", capa.features.String("SCardTransmit"), True),
-    ("mimikatz", "function=0x40105D", capa.features.String("ACR  > "), True),
-    ("mimikatz", "function=0x40105D", capa.features.String("nope"), False),
-    ("773290...", "function=0x140001140", capa.features.String(r"%s:\\OfficePackagesForWDAG"), True),
-    # insn/regex, issue #262
-    ("pma16-01", "function=0x4021B0", capa.features.Regex("HTTP/1.0"), True),
-    ("pma16-01", "function=0x4021B0", capa.features.Regex("www.practicalmalwareanalysis.com"), False),
-    # insn/string, pointer to string
-    ("mimikatz", "function=0x44EDEF", capa.features.String("INPUTEVENT"), True),
-    # insn/string, direct memory reference
-    ("mimikatz", "function=0x46D6CE", capa.features.String("(null)"), True),
-    # insn/bytes
-    ("mimikatz", "function=0x40105D", capa.features.Bytes("SCardControl".encode("utf-16le")), True),
-    ("mimikatz", "function=0x40105D", capa.features.Bytes("SCardTransmit".encode("utf-16le")), True),
-    ("mimikatz", "function=0x40105D", capa.features.Bytes("ACR  > ".encode("utf-16le")), True),
-    ("mimikatz", "function=0x40105D", capa.features.Bytes("nope".encode("ascii")), False),
-    # IDA features included byte sequences read from invalid memory, fixed in #409
-    ("mimikatz", "function=0x44570F", capa.features.Bytes(binascii.unhexlify("FF" * 256)), False),
-    # insn/bytes, pointer to bytes
-    ("mimikatz", "function=0x44EDEF", capa.features.Bytes("INPUTEVENT".encode("utf-16le")), True),
-    # insn/characteristic(nzxor)
-    ("mimikatz", "function=0x410DFC", capa.features.Characteristic("nzxor"), True),
-    ("mimikatz", "function=0x40105D", capa.features.Characteristic("nzxor"), False),
-    # insn/characteristic(nzxor): no security cookies
-    ("mimikatz", "function=0x46D534", capa.features.Characteristic("nzxor"), False),
-    # insn/characteristic(nzxor): xorps
-    # viv needs fixup to recognize function, see above
-    ("3b13b...", "function=0x10006860", capa.features.Characteristic("nzxor"), True),
-    # insn/characteristic(peb access)
-    ("kernel32-64", "function=0x1800017D0", capa.features.Characteristic("peb access"), True),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("peb access"), False),
-    # insn/characteristic(gs access)
-    ("kernel32-64", "function=0x180001068", capa.features.Characteristic("gs access"), True),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("gs access"), False),
-    # insn/characteristic(cross section flow)
-    ("a1982...", "function=0x4014D0", capa.features.Characteristic("cross section flow"), True),
-    # insn/characteristic(cross section flow): imports don't count
-    ("kernel32-64", "function=0x180001068", capa.features.Characteristic("cross section flow"), False),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("cross section flow"), False),
-    # insn/characteristic(recursive call)
-    ("39c05...", "function=0x10003100", capa.features.Characteristic("recursive call"), True),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("recursive call"), False),
-    # insn/characteristic(indirect call)
-    ("mimikatz", "function=0x4175FF", capa.features.Characteristic("indirect call"), True),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("indirect call"), False),
-    # insn/characteristic(calls from)
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("calls from"), True),
-    ("mimikatz", "function=0x4702FD", capa.features.Characteristic("calls from"), False),
-    # function/characteristic(calls to)
-    ("mimikatz", "function=0x40105D", capa.features.Characteristic("calls to"), True),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("calls to"), False),
-]
+FEATURE_PRESENCE_TESTS = sorted(
+    [
+        # file/characteristic("embedded pe")
+        ("pma12-04", "file", capa.features.common.Characteristic("embedded pe"), True),
+        # file/string
+        ("mimikatz", "file", capa.features.common.String("SCardControl"), True),
+        ("mimikatz", "file", capa.features.common.String("SCardTransmit"), True),
+        ("mimikatz", "file", capa.features.common.String("ACR  > "), True),
+        ("mimikatz", "file", capa.features.common.String("nope"), False),
+        # file/sections
+        ("mimikatz", "file", capa.features.file.Section(".text"), True),
+        ("mimikatz", "file", capa.features.file.Section(".nope"), False),
+        # IDA doesn't extract unmapped sections by default
+        # ("mimikatz", "file", capa.features.file.Section(".rsrc"), True),
+        # file/exports
+        ("kernel32", "file", capa.features.file.Export("BaseThreadInitThunk"), True),
+        ("kernel32", "file", capa.features.file.Export("lstrlenW"), True),
+        ("kernel32", "file", capa.features.file.Export("nope"), False),
+        # file/imports
+        ("mimikatz", "file", capa.features.file.Import("advapi32.CryptSetHashParam"), True),
+        ("mimikatz", "file", capa.features.file.Import("CryptSetHashParam"), True),
+        ("mimikatz", "file", capa.features.file.Import("kernel32.IsWow64Process"), True),
+        ("mimikatz", "file", capa.features.file.Import("msvcrt.exit"), True),
+        ("mimikatz", "file", capa.features.file.Import("cabinet.#11"), True),
+        ("mimikatz", "file", capa.features.file.Import("#11"), False),
+        ("mimikatz", "file", capa.features.file.Import("#nope"), False),
+        ("mimikatz", "file", capa.features.file.Import("nope"), False),
+        ("mimikatz", "file", capa.features.file.Import("advapi32.CryptAcquireContextW"), True),
+        ("mimikatz", "file", capa.features.file.Import("advapi32.CryptAcquireContext"), True),
+        ("mimikatz", "file", capa.features.file.Import("CryptAcquireContextW"), True),
+        ("mimikatz", "file", capa.features.file.Import("CryptAcquireContext"), True),
+        # function/characteristic(loop)
+        ("mimikatz", "function=0x401517", capa.features.common.Characteristic("loop"), True),
+        ("mimikatz", "function=0x401000", capa.features.common.Characteristic("loop"), False),
+        # bb/characteristic(tight loop)
+        ("mimikatz", "function=0x402EC4", capa.features.common.Characteristic("tight loop"), True),
+        ("mimikatz", "function=0x401000", capa.features.common.Characteristic("tight loop"), False),
+        # bb/characteristic(stack string)
+        ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("stack string"), True),
+        ("mimikatz", "function=0x401000", capa.features.common.Characteristic("stack string"), False),
+        # bb/characteristic(tight loop)
+        ("mimikatz", "function=0x402EC4,bb=0x402F8E", capa.features.common.Characteristic("tight loop"), True),
+        ("mimikatz", "function=0x401000,bb=0x401000", capa.features.common.Characteristic("tight loop"), False),
+        # insn/mnemonic
+        ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("push"), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("movzx"), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("xor"), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("in"), False),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Mnemonic("out"), False),
+        # insn/number
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0x3136B0), True),
+        # insn/number: stack adjustments
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xC), False),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0x10), False),
+        # insn/number: arch flavors
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF, arch=ARCH_X32), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Number(0xFF, arch=ARCH_X64), False),
+        # insn/offset
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x4), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0xC), True),
+        # insn/offset, issue #276
+        ("64d9f", "function=0x10001510,bb=0x100015B0", capa.features.insn.Offset(0x4000), True),
+        # insn/offset: stack references
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x8), False),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x10), False),
+        # insn/offset: negative
+        ("mimikatz", "function=0x4011FB", capa.features.insn.Offset(-0x1), True),
+        ("mimikatz", "function=0x4011FB", capa.features.insn.Offset(-0x2), True),
+        # insn/offset: arch flavors
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0, arch=ARCH_X32), True),
+        ("mimikatz", "function=0x40105D", capa.features.insn.Offset(0x0, arch=ARCH_X64), False),
+        # insn/api
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptAcquireContextW"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptAcquireContext"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptGenKey"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptImportKey"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.CryptDestroyKey"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptAcquireContextW"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptAcquireContext"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptGenKey"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptImportKey"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("CryptDestroyKey"), True),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("Nope"), False),
+        ("mimikatz", "function=0x403BAC", capa.features.insn.API("advapi32.Nope"), False),
+        # insn/api: thunk
+        ("mimikatz", "function=0x4556E5", capa.features.insn.API("advapi32.LsaQueryInformationPolicy"), True),
+        ("mimikatz", "function=0x4556E5", capa.features.insn.API("LsaQueryInformationPolicy"), True),
+        # insn/api: x64
+        (
+            "kernel32-64",
+            "function=0x180001010",
+            capa.features.insn.API("RtlVirtualUnwind"),
+            True,
+        ),
+        ("kernel32-64", "function=0x180001010", capa.features.insn.API("RtlVirtualUnwind"), True),
+        # insn/api: x64 thunk
+        (
+            "kernel32-64",
+            "function=0x1800202B0",
+            capa.features.insn.API("RtlCaptureContext"),
+            True,
+        ),
+        ("kernel32-64", "function=0x1800202B0", capa.features.insn.API("RtlCaptureContext"), True),
+        # insn/api: x64 nested thunk
+        ("al-khaser x64", "function=0x14004B4F0", capa.features.insn.API("__vcrt_GetModuleHandle"), True),
+        # insn/api: call via jmp
+        ("mimikatz", "function=0x40B3C6", capa.features.insn.API("LocalFree"), True),
+        ("c91887...", "function=0x40156F", capa.features.insn.API("CloseClipboard"), True),
+        # TODO ignore thunk functions that call via jmp?
+        # insn/api: resolve indirect calls
+        ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.CreatePipe"), True),
+        ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.SetHandleInformation"), True),
+        ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.CloseHandle"), True),
+        ("c91887...", "function=0x401A77", capa.features.insn.API("kernel32.WriteFile"), True),
+        # insn/string
+        ("mimikatz", "function=0x40105D", capa.features.common.String("SCardControl"), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.String("SCardTransmit"), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.String("ACR  > "), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.String("nope"), False),
+        ("773290...", "function=0x140001140", capa.features.common.String(r"%s:\\OfficePackagesForWDAG"), True),
+        # insn/regex, issue #262
+        ("pma16-01", "function=0x4021B0", capa.features.common.Regex("HTTP/1.0"), True),
+        ("pma16-01", "function=0x4021B0", capa.features.common.Regex("www.practicalmalwareanalysis.com"), False),
+        # insn/string, pointer to string
+        ("mimikatz", "function=0x44EDEF", capa.features.common.String("INPUTEVENT"), True),
+        # insn/string, direct memory reference
+        ("mimikatz", "function=0x46D6CE", capa.features.common.String("(null)"), True),
+        # insn/bytes
+        ("mimikatz", "function=0x40105D", capa.features.common.Bytes("SCardControl".encode("utf-16le")), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.Bytes("SCardTransmit".encode("utf-16le")), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.Bytes("ACR  > ".encode("utf-16le")), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.Bytes("nope".encode("ascii")), False),
+        # IDA features included byte sequences read from invalid memory, fixed in #409
+        ("mimikatz", "function=0x44570F", capa.features.common.Bytes(binascii.unhexlify("FF" * 256)), False),
+        # insn/bytes, pointer to bytes
+        ("mimikatz", "function=0x44EDEF", capa.features.common.Bytes("INPUTEVENT".encode("utf-16le")), True),
+        # insn/characteristic(nzxor)
+        ("mimikatz", "function=0x410DFC", capa.features.common.Characteristic("nzxor"), True),
+        ("mimikatz", "function=0x40105D", capa.features.common.Characteristic("nzxor"), False),
+        # insn/characteristic(nzxor): no security cookies
+        ("mimikatz", "function=0x46D534", capa.features.common.Characteristic("nzxor"), False),
+        # insn/characteristic(nzxor): xorps
+        # viv needs fixup to recognize function, see above
+        ("3b13b...", "function=0x10006860", capa.features.common.Characteristic("nzxor"), True),
+        # insn/characteristic(peb access)
+        ("kernel32-64", "function=0x1800017D0", capa.features.common.Characteristic("peb access"), True),
+        ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("peb access"), False),
+        # insn/characteristic(gs access)
+        ("kernel32-64", "function=0x180001068", capa.features.common.Characteristic("gs access"), True),
+        ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("gs access"), False),
+        # insn/characteristic(cross section flow)
+        ("a1982...", "function=0x4014D0", capa.features.common.Characteristic("cross section flow"), True),
+        # insn/characteristic(cross section flow): imports don't count
+        ("kernel32-64", "function=0x180001068", capa.features.common.Characteristic("cross section flow"), False),
+        ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("cross section flow"), False),
+        # insn/characteristic(recursive call)
+        ("mimikatz", "function=0x40640e", capa.features.common.Characteristic("recursive call"), True),
+        # before this we used ambiguous (0x4556E5, False), which has a data reference / indirect recursive call, see #386
+        ("mimikatz", "function=0x4175FF", capa.features.common.Characteristic("recursive call"), False),
+        # insn/characteristic(indirect call)
+        ("mimikatz", "function=0x4175FF", capa.features.common.Characteristic("indirect call"), True),
+        ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("indirect call"), False),
+        # insn/characteristic(calls from)
+        ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("calls from"), True),
+        ("mimikatz", "function=0x4702FD", capa.features.common.Characteristic("calls from"), False),
+        # function/characteristic(calls to)
+        ("mimikatz", "function=0x40105D", capa.features.common.Characteristic("calls to"), True),
+        # before this we used ambiguous (0x4556E5, False), which has a data reference / indirect recursive call, see #386
+        ("mimikatz", "function=0x456BB9", capa.features.common.Characteristic("calls to"), False),
+        # file/function-name
+        ("pma16-01", "file", capa.features.file.FunctionName("__aulldiv"), True),
+    ],
+    # order tests by (file, item)
+    # so that our LRU cache is most effective.
+    key=lambda t: (t[0], t[1]),
+)
 
 FEATURE_PRESENCE_TESTS_IDA = [
     # file/imports
@@ -490,10 +513,10 @@ FEATURE_PRESENCE_TESTS_IDA = [
 
 FEATURE_COUNT_TESTS = [
     ("mimikatz", "function=0x40E5C2", capa.features.basicblock.BasicBlock(), 7),
-    ("mimikatz", "function=0x4702FD", capa.features.Characteristic("calls from"), 0),
-    ("mimikatz", "function=0x40E5C2", capa.features.Characteristic("calls from"), 3),
-    ("mimikatz", "function=0x4556E5", capa.features.Characteristic("calls to"), 0),
-    ("mimikatz", "function=0x40B1F1", capa.features.Characteristic("calls to"), 3),
+    ("mimikatz", "function=0x4702FD", capa.features.common.Characteristic("calls from"), 0),
+    ("mimikatz", "function=0x40E5C2", capa.features.common.Characteristic("calls from"), 3),
+    ("mimikatz", "function=0x4556E5", capa.features.common.Characteristic("calls to"), 0),
+    ("mimikatz", "function=0x40B1F1", capa.features.common.Characteristic("calls to"), 3),
 ]
 
 

tests/test_engine.py

@@ -10,6 +10,8 @@ import textwrap
 
 import capa.rules
 import capa.engine
+import capa.features.insn
+import capa.features.common
 from capa.engine import *
 from capa.features import *
 from capa.features.insn import *
@@ -233,7 +235,7 @@ def test_match_adds_matched_rule_feature():
     )
     r = capa.rules.Rule.from_yaml(rule)
     features, matches = capa.engine.match([r], {capa.features.insn.Number(100): {1}}, 0x0)
-    assert capa.features.MatchedRule("test rule") in features
+    assert capa.features.common.MatchedRule("test rule") in features
 
 
 def test_match_matched_rules():
@@ -264,22 +266,22 @@ def test_match_matched_rules():
     ]
 
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
+        capa.rules.topologically_order_rules(rules),
         {capa.features.insn.Number(100): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule1") in features
-    assert capa.features.MatchedRule("test rule2") in features
+    assert capa.features.common.MatchedRule("test rule1") in features
+    assert capa.features.common.MatchedRule("test rule2") in features
 
     # the ordering of the rules must not matter,
     # the engine should match rules in an appropriate order.
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(reversed(rules)),
+        capa.rules.topologically_order_rules(reversed(rules)),
         {capa.features.insn.Number(100): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule1") in features
-    assert capa.features.MatchedRule("test rule2") in features
+    assert capa.features.common.MatchedRule("test rule1") in features
+    assert capa.features.common.MatchedRule("test rule2") in features
 
 
 def test_regex():
@@ -322,34 +324,34 @@ def test_regex():
         ),
     ]
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
+        capa.rules.topologically_order_rules(rules),
         {capa.features.insn.Number(100): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule") not in features
+    assert capa.features.common.MatchedRule("test rule") not in features
 
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
-        {capa.features.String("aaaa"): {1}},
+        capa.rules.topologically_order_rules(rules),
+        {capa.features.common.String("aaaa"): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule") not in features
+    assert capa.features.common.MatchedRule("test rule") not in features
 
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
-        {capa.features.String("aBBBBa"): {1}},
+        capa.rules.topologically_order_rules(rules),
+        {capa.features.common.String("aBBBBa"): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule") not in features
+    assert capa.features.common.MatchedRule("test rule") not in features
 
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
-        {capa.features.String("abbbba"): {1}},
+        capa.rules.topologically_order_rules(rules),
+        {capa.features.common.String("abbbba"): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule") in features
-    assert capa.features.MatchedRule("rule with implied wildcards") in features
-    assert capa.features.MatchedRule("rule with anchor") not in features
+    assert capa.features.common.MatchedRule("test rule") in features
+    assert capa.features.common.MatchedRule("rule with implied wildcards") in features
+    assert capa.features.common.MatchedRule("rule with anchor") not in features
 
 
 def test_regex_ignorecase():
@@ -368,11 +370,11 @@ def test_regex_ignorecase():
         ),
     ]
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
-        {capa.features.String("aBBBBa"): {1}},
+        capa.rules.topologically_order_rules(rules),
+        {capa.features.common.String("aBBBBa"): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule") in features
+    assert capa.features.common.MatchedRule("test rule") in features
 
 
 def test_regex_complex():
@@ -391,11 +393,11 @@ def test_regex_complex():
         ),
     ]
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
-        {capa.features.String(r"Hardware\Key\key with spaces\some value"): {1}},
+        capa.rules.topologically_order_rules(rules),
+        {capa.features.common.String(r"Hardware\Key\key with spaces\some value"): {1}},
         0x0,
     )
-    assert capa.features.MatchedRule("test rule") in features
+    assert capa.features.common.MatchedRule("test rule") in features
 
 
 def test_match_namespace():
@@ -449,19 +451,19 @@ def test_match_namespace():
     ]
 
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
+        capa.rules.topologically_order_rules(rules),
         {capa.features.insn.API("CreateFile"): {1}},
         0x0,
     )
     assert "CreateFile API" in matches
     assert "file-create" in matches
     assert "filesystem-any" in matches
-    assert capa.features.MatchedRule("file") in features
-    assert capa.features.MatchedRule("file/create") in features
-    assert capa.features.MatchedRule("file/create/CreateFile") in features
+    assert capa.features.common.MatchedRule("file") in features
+    assert capa.features.common.MatchedRule("file/create") in features
+    assert capa.features.common.MatchedRule("file/create/CreateFile") in features
 
     features, matches = capa.engine.match(
-        capa.engine.topologically_order_rules(rules),
+        capa.rules.topologically_order_rules(rules),
         {capa.features.insn.API("WriteFile"): {1}},
         0x0,
     )
@@ -472,11 +474,11 @@ def test_match_namespace():
 
 def test_render_number():
     assert str(capa.features.insn.Number(1)) == "number(0x1)"
-    assert str(capa.features.insn.Number(1, arch=ARCH_X32)) == "number/x32(0x1)"
-    assert str(capa.features.insn.Number(1, arch=ARCH_X64)) == "number/x64(0x1)"
+    assert str(capa.features.insn.Number(1, arch=capa.features.common.ARCH_X32)) == "number/x32(0x1)"
+    assert str(capa.features.insn.Number(1, arch=capa.features.common.ARCH_X64)) == "number/x64(0x1)"
 
 
 def test_render_offset():
     assert str(capa.features.insn.Offset(1)) == "offset(0x1)"
-    assert str(capa.features.insn.Offset(1, arch=ARCH_X32)) == "offset/x32(0x1)"
-    assert str(capa.features.insn.Offset(1, arch=ARCH_X64)) == "offset/x64(0x1)"
+    assert str(capa.features.insn.Offset(1, arch=capa.features.common.ARCH_X32)) == "offset/x32(0x1)"
+    assert str(capa.features.insn.Offset(1, arch=capa.features.common.ARCH_X64)) == "offset/x64(0x1)"

tests/test_freeze.py

@@ -5,40 +5,41 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-import sys
 import textwrap
 
-import pytest
 from fixtures import *
 
 import capa.main
+import capa.rules
 import capa.helpers
-import capa.features
+import capa.features.file
 import capa.features.insn
+import capa.features.common
 import capa.features.freeze
-import capa.features.extractors
+import capa.features.basicblock
+import capa.features.extractors.base_extractor
 
-EXTRACTOR = capa.features.extractors.NullFeatureExtractor(
+EXTRACTOR = capa.features.extractors.base_extractor.NullFeatureExtractor(
     {
         "base address": 0x401000,
         "file features": [
-            (0x402345, capa.features.Characteristic("embedded pe")),
+            (0x402345, capa.features.common.Characteristic("embedded pe")),
         ],
         "functions": {
             0x401000: {
                 "features": [
-                    (0x401000, capa.features.Characteristic("indirect call")),
+                    (0x401000, capa.features.common.Characteristic("indirect call")),
                 ],
                 "basic blocks": {
                     0x401000: {
                         "features": [
-                            (0x401000, capa.features.Characteristic("tight loop")),
+                            (0x401000, capa.features.common.Characteristic("tight loop")),
                         ],
                         "instructions": {
                             0x401000: {
                                 "features": [
                                     (0x401000, capa.features.insn.Mnemonic("xor")),
-                                    (0x401000, capa.features.Characteristic("nzxor")),
+                                    (0x401000, capa.features.common.Characteristic("nzxor")),
                                 ],
                             },
                             0x401002: {
@@ -112,34 +113,27 @@ def compare_extractors_viv_null(viv_ext, null_ext):
     and NullFeatureExtractor returns ints
 
     args:
-      viv_ext (capa.features.extractors.viv.VivisectFeatureExtractor)
+      viv_ext (capa.features.extractors.viv.extractor.VivisectFeatureExtractor)
       null_ext (capa.features.extractors.NullFeatureExtractor)
     """
     assert list(viv_ext.extract_file_features()) == list(null_ext.extract_file_features())
-    assert list(map(to_int, viv_ext.get_functions())) == list(null_ext.get_functions())
+    assert list(map(int, viv_ext.get_functions())) == list(null_ext.get_functions())
     for f in viv_ext.get_functions():
-        assert list(map(to_int, viv_ext.get_basic_blocks(f))) == list(null_ext.get_basic_blocks(to_int(f)))
-        assert list(viv_ext.extract_function_features(f)) == list(null_ext.extract_function_features(to_int(f)))
+        assert list(map(int, viv_ext.get_basic_blocks(f))) == list(null_ext.get_basic_blocks(int(f)))
+        assert list(viv_ext.extract_function_features(f)) == list(null_ext.extract_function_features(int(f)))
 
         for bb in viv_ext.get_basic_blocks(f):
-            assert list(map(to_int, viv_ext.get_instructions(f, bb))) == list(
-                null_ext.get_instructions(to_int(f), to_int(bb))
-            )
+            assert list(map(int, viv_ext.get_instructions(f, bb))) == list(null_ext.get_instructions(int(f), int(bb)))
             assert list(viv_ext.extract_basic_block_features(f, bb)) == list(
-                null_ext.extract_basic_block_features(to_int(f), to_int(bb))
+                null_ext.extract_basic_block_features(int(f), int(bb))
             )
 
             for insn in viv_ext.get_instructions(f, bb):
                 assert list(viv_ext.extract_insn_features(f, bb, insn)) == list(
-                    null_ext.extract_insn_features(to_int(f), to_int(bb), to_int(insn))
+                    null_ext.extract_insn_features(int(f), int(bb), int(insn))
                 )
 
 
-def to_int(o):
-    """helper to get int value of extractor items"""
-    return capa.helpers.oint(o)
-
-
 def test_freeze_s_roundtrip():
     load = capa.features.freeze.loads
     dump = capa.features.freeze.dumps
@@ -162,19 +156,18 @@ def roundtrip_feature(feature):
 
 def test_serialize_features():
     roundtrip_feature(capa.features.insn.API("advapi32.CryptAcquireContextW"))
-    roundtrip_feature(capa.features.String("SCardControl"))
+    roundtrip_feature(capa.features.common.String("SCardControl"))
     roundtrip_feature(capa.features.insn.Number(0xFF))
     roundtrip_feature(capa.features.insn.Offset(0x0))
     roundtrip_feature(capa.features.insn.Mnemonic("push"))
     roundtrip_feature(capa.features.file.Section(".rsrc"))
-    roundtrip_feature(capa.features.Characteristic("tight loop"))
+    roundtrip_feature(capa.features.common.Characteristic("tight loop"))
     roundtrip_feature(capa.features.basicblock.BasicBlock())
     roundtrip_feature(capa.features.file.Export("BaseThreadInitThunk"))
     roundtrip_feature(capa.features.file.Import("kernel32.IsWow64Process"))
     roundtrip_feature(capa.features.file.Import("#11"))
 
 
-@pytest.mark.xfail(sys.version_info >= (3, 0), reason="vivsect only works on py2")
 def test_freeze_sample(tmpdir, z9324d_extractor):
     # tmpdir fixture handles cleanup
     o = tmpdir.mkdir("capa").join("test.frz").strpath
@@ -182,7 +175,6 @@ def test_freeze_sample(tmpdir, z9324d_extractor):
     assert capa.features.freeze.main([path, o, "-v"]) == 0
 
 
-@pytest.mark.xfail(sys.version_info >= (3, 0), reason="vivsect only works on py2")
 def test_freeze_load_sample(tmpdir, z9324d_extractor):
     o = tmpdir.mkdir("capa").join("test.frz")
 

tests/test_function_id.py

@@ -0,0 +1,27 @@
+from fixtures import get_function, pma16_01_extractor, extract_function_features
+
+import capa.features.insn
+
+
+def test_function_id_simple_match(pma16_01_extractor):
+    assert pma16_01_extractor.is_library_function(0x407490) == True
+    assert pma16_01_extractor.get_function_name(0x407490) == "__aulldiv"
+
+
+def test_function_id_gz_pat(pma16_01_extractor):
+    # aullrem is stored in `test_aullrem.pat.gz`
+    assert pma16_01_extractor.is_library_function(0x407500) == True
+    assert pma16_01_extractor.get_function_name(0x407500) == "__aullrem"
+
+
+def test_function_id_complex_match(pma16_01_extractor):
+    # 0x405714 is __spawnlp which requires recursive match of __spawnvp at 0x407FAB
+    # (and __spawnvpe at 0x409DE8)
+    assert pma16_01_extractor.is_library_function(0x405714) == True
+    assert pma16_01_extractor.get_function_name(0x405714) == "__spawnlp"
+
+
+def test_function_id_api_feature(pma16_01_extractor):
+    f = get_function(pma16_01_extractor, 0x404548)
+    features = extract_function_features(pma16_01_extractor, f)
+    assert capa.features.insn.API("__aulldiv") in features

tests/test_helpers.py

@@ -12,12 +12,10 @@ from capa.features.extractors import helpers
 
 
 def test_all_zeros():
-    # Python 2: <str>
-    # Python 3: <bytes>
     a = b"\x00\x00\x00\x00"
-    b = codecs.decode("00000000", "hex")
+    b = codecs.decode(b"00000000", "hex")
     c = b"\x01\x00\x00\x00"
-    d = codecs.decode("01000000", "hex")
+    d = codecs.decode(b"01000000", "hex")
     assert helpers.all_zeros(a) is True
     assert helpers.all_zeros(b) is True
     assert helpers.all_zeros(c) is False

tests/test_ida_features.py

@@ -1,104 +1,105 @@
-# run this script from within IDA with ./tests/data/mimikatz.exe open
-import sys
-import logging
-import os.path
-import binascii
-import traceback
-
-import pytest
-
-try:
-    sys.path.append(os.path.dirname(__file__))
-    from fixtures import *
-finally:
-    sys.path.pop()
-
-
-logger = logging.getLogger("test_ida_features")
-
-
-def check_input_file(wanted):
-    import idautils
-
-    # some versions (7.4) of IDA return a truncated version of the MD5.
-    # https://github.com/idapython/bin/issues/11
-    try:
-        found = idautils.GetInputFileMD5()[:31].decode("ascii").lower()
-    except UnicodeDecodeError:
-        # in IDA 7.5 or so, GetInputFileMD5 started returning raw binary
-        # rather than the hex digest
-        found = binascii.hexlify(idautils.GetInputFileMD5()[:15]).decode("ascii").lower()
-
-    if not wanted.startswith(found):
-        raise RuntimeError("please run the tests against sample with MD5: `%s`" % (wanted))
-
-
-def get_ida_extractor(_path):
-    check_input_file("5f66b82558ca92e54e77f216ef4c066c")
-
-    # have to import import this inline so pytest doesn't bail outside of IDA
-    import capa.features.extractors.ida
-
-    return capa.features.extractors.ida.IdaFeatureExtractor()
-
-
-@pytest.mark.skip(reason="IDA Pro tests must be run within IDA")
-def test_ida_features():
-    for (sample, scope, feature, expected) in FEATURE_PRESENCE_TESTS + FEATURE_PRESENCE_TESTS_IDA:
-        id = make_test_id((sample, scope, feature, expected))
-
-        try:
-            check_input_file(get_sample_md5_by_name(sample))
-        except RuntimeError:
-            print("SKIP %s" % (id))
-            continue
-
-        scope = resolve_scope(scope)
-        sample = resolve_sample(sample)
-
-        try:
-            do_test_feature_presence(get_ida_extractor, sample, scope, feature, expected)
-        except Exception as e:
-            print("FAIL %s" % (id))
-            traceback.print_exc()
-        else:
-            print("OK   %s" % (id))
-
-
-@pytest.mark.skip(reason="IDA Pro tests must be run within IDA")
-def test_ida_feature_counts():
-    for (sample, scope, feature, expected) in FEATURE_COUNT_TESTS:
-        id = make_test_id((sample, scope, feature, expected))
-
-        try:
-            check_input_file(get_sample_md5_by_name(sample))
-        except RuntimeError:
-            print("SKIP %s" % (id))
-            continue
-
-        scope = resolve_scope(scope)
-        sample = resolve_sample(sample)
-
-        try:
-            do_test_feature_count(get_ida_extractor, sample, scope, feature, expected)
-        except Exception as e:
-            print("FAIL %s" % (id))
-            traceback.print_exc()
-        else:
-            print("OK   %s" % (id))
-
-
-if __name__ == "__main__":
-    print("-" * 80)
-
-    # invoke all functions in this module that start with `test_`
-    for name in dir(sys.modules[__name__]):
-        if not name.startswith("test_"):
-            continue
-
-        test = getattr(sys.modules[__name__], name)
-        logger.debug("invoking test: %s", name)
-        sys.stderr.flush()
-        test()
-
-    print("DONE")
+# run this script from within IDA with ./tests/data/mimikatz.exe open
+import sys
+import logging
+import os.path
+import binascii
+import traceback
+
+import pytest
+
+try:
+    sys.path.append(os.path.dirname(__file__))
+    import fixtures
+    from fixtures import *
+finally:
+    sys.path.pop()
+
+
+logger = logging.getLogger("test_ida_features")
+
+
+def check_input_file(wanted):
+    import idautils
+
+    # some versions (7.4) of IDA return a truncated version of the MD5.
+    # https://github.com/idapython/bin/issues/11
+    try:
+        found = idautils.GetInputFileMD5()[:31].decode("ascii").lower()
+    except UnicodeDecodeError:
+        # in IDA 7.5 or so, GetInputFileMD5 started returning raw binary
+        # rather than the hex digest
+        found = binascii.hexlify(idautils.GetInputFileMD5()[:15]).decode("ascii").lower()
+
+    if not wanted.startswith(found):
+        raise RuntimeError("please run the tests against sample with MD5: `%s`" % (wanted))
+
+
+def get_ida_extractor(_path):
+    check_input_file("5f66b82558ca92e54e77f216ef4c066c")
+
+    # have to import import this inline so pytest doesn't bail outside of IDA
+    import capa.features.extractors.ida.extractor
+
+    return capa.features.extractors.ida.extractor.IdaFeatureExtractor()
+
+
+@pytest.mark.skip(reason="IDA Pro tests must be run within IDA")
+def test_ida_features():
+    for (sample, scope, feature, expected) in fixtures.FEATURE_PRESENCE_TESTS + fixtures.FEATURE_PRESENCE_TESTS_IDA:
+        id = fixtures.make_test_id((sample, scope, feature, expected))
+
+        try:
+            check_input_file(fixtures.get_sample_md5_by_name(sample))
+        except RuntimeError:
+            print("SKIP %s" % (id))
+            continue
+
+        scope = fixtures.resolve_scope(scope)
+        sample = fixtures.resolve_sample(sample)
+
+        try:
+            fixtures.do_test_feature_presence(get_ida_extractor, sample, scope, feature, expected)
+        except Exception as e:
+            print("FAIL %s" % (id))
+            traceback.print_exc()
+        else:
+            print("OK   %s" % (id))
+
+
+@pytest.mark.skip(reason="IDA Pro tests must be run within IDA")
+def test_ida_feature_counts():
+    for (sample, scope, feature, expected) in fixtures.FEATURE_COUNT_TESTS:
+        id = fixtures.make_test_id((sample, scope, feature, expected))
+
+        try:
+            check_input_file(fixtures.get_sample_md5_by_name(sample))
+        except RuntimeError:
+            print("SKIP %s" % (id))
+            continue
+
+        scope = fixtures.resolve_scope(scope)
+        sample = fixtures.resolve_sample(sample)
+
+        try:
+            fixtures.do_test_feature_count(get_ida_extractor, sample, scope, feature, expected)
+        except Exception as e:
+            print("FAIL %s" % (id))
+            traceback.print_exc()
+        else:
+            print("OK   %s" % (id))
+
+
+if __name__ == "__main__":
+    print("-" * 80)
+
+    # invoke all functions in this module that start with `test_`
+    for name in dir(sys.modules[__name__]):
+        if not name.startswith("test_"):
+            continue
+
+        test = getattr(sys.modules[__name__], name)
+        logger.debug("invoking test: %s", name)
+        sys.stderr.flush()
+        test()
+
+    print("DONE")

tests/test_main.py

@@ -6,11 +6,10 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-import sys
 import json
 import textwrap
 
-import pytest
+import fixtures
 from fixtures import *
 
 import capa.main
@@ -26,6 +25,7 @@ def test_main(z9324d_extractor):
     assert capa.main.main([path, "-vv"]) == 0
     assert capa.main.main([path, "-v"]) == 0
     assert capa.main.main([path, "-j"]) == 0
+    assert capa.main.main([path, "-q"]) == 0
     assert capa.main.main([path]) == 0
 
 
@@ -58,10 +58,6 @@ def test_main_single_rule(z9324d_extractor, tmpdir):
 
 
 def test_main_non_ascii_filename(pingtaest_extractor, tmpdir, capsys):
-    # on py2.7, need to be careful about str (which can hold bytes)
-    #  vs unicode (which is only unicode characters).
-    # on py3, this should not be needed.
-    #
     # here we print a string with unicode characters in it
     # (specifically, a byte string with utf-8 bytes in it, see file encoding)
     assert capa.main.main(["-q", pingtaest_extractor.path]) == 0
@@ -69,20 +65,14 @@ def test_main_non_ascii_filename(pingtaest_extractor, tmpdir, capsys):
     std = capsys.readouterr()
     # but here, we have to use a unicode instance,
     # because capsys has decoded the output for us.
-    if sys.version_info >= (3, 0):
-        assert pingtaest_extractor.path in std.out
-    else:
-        assert pingtaest_extractor.path.decode("utf-8") in std.out
+    assert pingtaest_extractor.path in std.out
 
 
 def test_main_non_ascii_filename_nonexistent(tmpdir, caplog):
     NON_ASCII_FILENAME = "täst_not_there.exe"
     assert capa.main.main(["-q", NON_ASCII_FILENAME]) == -1
 
-    if sys.version_info >= (3, 0):
-        assert NON_ASCII_FILENAME in caplog.text
-    else:
-        assert NON_ASCII_FILENAME.decode("utf-8") in caplog.text
+    assert NON_ASCII_FILENAME in caplog.text
 
 
 def test_main_shellcode(z499c2_extractor):
@@ -90,7 +80,10 @@ def test_main_shellcode(z499c2_extractor):
     assert capa.main.main([path, "-vv", "-f", "sc32"]) == 0
     assert capa.main.main([path, "-v", "-f", "sc32"]) == 0
     assert capa.main.main([path, "-j", "-f", "sc32"]) == 0
+    assert capa.main.main([path, "-q", "-f", "sc32"]) == 0
     assert capa.main.main([path, "-f", "sc32"]) == 0
+    # auto detect shellcode based on file extension
+    assert capa.main.main([path]) == 0
 
 
 def test_ruleset():
@@ -368,18 +361,17 @@ def test_not_render_rules_also_matched(z9324d_extractor, capsys):
     assert "create TCP socket" in std.out
 
 
-# It tests main works with different backends
 def test_backend_option(capsys):
-    if sys.version_info > (3, 0):
-        path = get_data_path_by_name("pma16-01")
-        assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_VIV]) == 0
-        std = capsys.readouterr()
-        std_json = json.loads(std.out)
-        assert std_json["meta"]["analysis"]["extractor"] == "VivisectFeatureExtractor"
-        assert len(std_json["rules"]) > 0
+    # tests that main works with different backends
+    path = fixtures.get_data_path_by_name("pma16-01")
+    assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_VIV]) == 0
+    std = capsys.readouterr()
+    std_json = json.loads(std.out)
+    assert std_json["meta"]["analysis"]["extractor"] == "VivisectFeatureExtractor"
+    assert len(std_json["rules"]) > 0
 
-        assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_SMDA]) == 0
-        std = capsys.readouterr()
-        std_json = json.loads(std.out)
-        assert std_json["meta"]["analysis"]["extractor"] == "SmdaFeatureExtractor"
-        assert len(std_json["rules"]) > 0
+    assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_SMDA]) == 0
+    std = capsys.readouterr()
+    std_json = json.loads(std.out)
+    assert std_json["meta"]["analysis"]["extractor"] == "SmdaFeatureExtractor"
+    assert len(std_json["rules"]) > 0

tests/test_pefile_features.py

@@ -0,0 +1,28 @@
+# Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at: [package root]/LICENSE.txt
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+#  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+import pytest
+import fixtures
+from fixtures import *
+from fixtures import parametrize
+
+import capa.features.file
+
+
+@parametrize(
+    "sample,scope,feature,expected",
+    fixtures.FEATURE_PRESENCE_TESTS,
+    indirect=["sample", "scope"],
+)
+def test_pefile_features(sample, scope, feature, expected):
+    if scope.__name__ != "file":
+        pytest.xfail("pefile only extract file scope features")
+
+    if isinstance(feature, capa.features.file.FunctionName):
+        pytest.xfail("pefile only doesn't extract function names")
+
+    fixtures.do_test_feature_presence(fixtures.get_pefile_extractor, sample, scope, feature, expected)

tests/test_render.py

@@ -0,0 +1,71 @@
+import textwrap
+
+import capa.rules
+import capa.render.utils
+import capa.render.result_document
+
+
+def test_render_meta_attack():
+    # Persistence::Boot or Logon Autostart Execution::Registry Run Keys / Startup Folder [T1547.001]
+    id = "T1543.003"
+    tactic = "Persistence"
+    technique = "Create or Modify System Process"
+    subtechnique = "Windows Service"
+    canonical = "{:s}::{:s}::{:s} [{:s}]".format(tactic, technique, subtechnique, id)
+
+    rule = textwrap.dedent(
+        """
+        rule:
+          meta:
+            name: test rule
+            att&ck:
+              - {:s}
+          features:
+            - number: 1
+        """.format(
+            canonical
+        )
+    )
+    r = capa.rules.Rule.from_yaml(rule)
+    rule_meta = capa.render.result_document.convert_meta_to_result_document(r.meta)
+    attack = rule_meta["att&ck"][0]
+
+    assert attack["id"] == id
+    assert attack["tactic"] == tactic
+    assert attack["technique"] == technique
+    assert attack["subtechnique"] == subtechnique
+
+    assert capa.render.utils.format_parts_id(attack) == canonical
+
+
+def test_render_meta_mbc():
+    # Defense Evasion::Disable or Evade Security Tools::Heavens Gate [F0004.008]
+    id = "F0004.008"
+    objective = "Defense Evasion"
+    behavior = "Disable or Evade Security Tools"
+    method = "Heavens Gate"
+    canonical = "{:s}::{:s}::{:s} [{:s}]".format(objective, behavior, method, id)
+
+    rule = textwrap.dedent(
+        """
+        rule:
+          meta:
+            name: test rule
+            mbc:
+              - {:s}
+          features:
+            - number: 1
+        """.format(
+            canonical
+        )
+    )
+    r = capa.rules.Rule.from_yaml(rule)
+    rule_meta = capa.render.result_document.convert_meta_to_result_document(r.meta)
+    attack = rule_meta["mbc"][0]
+
+    assert attack["id"] == id
+    assert attack["objective"] == objective
+    assert attack["behavior"] == behavior
+    assert attack["method"] == method
+
+    assert capa.render.utils.format_parts_id(attack) == canonical