tests: idalib: use 9.1 instead of 9.0 as min ver

9.0 doesn't support disabling lumina (or loading resources, for that
matter, too)

.github/pyinstaller/pyinstaller.spec

@@ -74,7 +74,6 @@ a = Analysis(
         # only be installed locally.
         "binaryninja",
         "ida",
-        "ghidra",
         # remove once https://github.com/mandiant/capa/issues/2681 has
         # been addressed by PyInstaller
         "pkg_resources",

.github/workflows/build.yml

@@ -28,11 +28,6 @@ jobs:
             artifact_name: capa
             asset_name: linux
             python_version: '3.10'
-            # for Ghidra
-            java-version: '21'
-            ghidra-version: '12.0'
-            public-version: 'PUBLIC_20251205'
-            ghidra-sha256: 'af43e8cfb2fa4490cf6020c3a2bde25c159d83f45236a0542688a024e8fc1941'
           - os: ubuntu-22.04-arm
             artifact_name: capa
             asset_name: linux-arm64
@@ -111,24 +106,6 @@ jobs:
         run: |
           7z e "tests/data/dynamic/cape/v2.2/d46900384c78863420fb3e297d0a2f743cd2b6b3f7f82bf64059a168e07aceb7.json.gz"
           dist/capa -d "d46900384c78863420fb3e297d0a2f743cd2b6b3f7f82bf64059a168e07aceb7.json"
-      - name: Set up Java ${{ matrix.java-version }}
-        if: matrix.os == 'ubuntu-22.04' && matrix.python_version == '3.10'
-        uses: actions/setup-java@387ac29b308b003ca37ba93a6cab5eb57c8f5f93 # v4.0.0
-        with:
-          distribution: 'temurin'
-          java-version: ${{ matrix.java-version }}
-      - name: Install Ghidra ${{ matrix.ghidra-version }}
-        if: matrix.os == 'ubuntu-22.04' && matrix.python_version == '3.10'
-        run: |
-          mkdir ./.github/ghidra
-          wget "https://github.com/NationalSecurityAgency/ghidra/releases/download/Ghidra_${{ matrix.ghidra-version }}_build/ghidra_${{ matrix.ghidra-version }}_${{ matrix.public-version }}.zip" -O ./.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC.zip
-          echo "${{ matrix.ghidra-sha256 }} ./.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC.zip" | sha256sum -c -
-          unzip .github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC.zip -d .github/ghidra/
-      - name: Does it run (Ghidra)?
-        if: matrix.os == 'ubuntu-22.04' && matrix.python_version == '3.10'
-        env:
-          GHIDRA_INSTALL_DIR: ${{ github.workspace }}/.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC
-        run: dist/capa -b ghidra -d "tests/data/Practical Malware Analysis Lab 01-01.dll_"
       - uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3 # v4.3.1
         with:
           name: ${{ matrix.asset_name }}

.github/workflows/tests.yml

@@ -172,10 +172,10 @@ jobs:
       fail-fast: false
       matrix:
         python-version: ["3.10", "3.13"]
-        java-version: ["21"]
-        ghidra-version: ["12.0"]
-        public-version: ["PUBLIC_20251205"] # for ghidra releases
-        ghidra-sha256: ['af43e8cfb2fa4490cf6020c3a2bde25c159d83f45236a0542688a024e8fc1941']
+        java-version: ["17"]
+        ghidra-version: ["11.0.1"]
+        public-version: ["PUBLIC_20240130"] # for ghidra releases
+        ghidrathon-version: ["4.0.0"] 
     steps:
     - name: Checkout capa with submodules
       uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -194,15 +194,75 @@ jobs:
       run: |
         mkdir ./.github/ghidra
         wget "https://github.com/NationalSecurityAgency/ghidra/releases/download/Ghidra_${{ matrix.ghidra-version }}_build/ghidra_${{ matrix.ghidra-version }}_${{ matrix.public-version }}.zip" -O ./.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC.zip
-        echo "${{ matrix.ghidra-sha256 }} ./.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC.zip" | sha256sum -c -
         unzip .github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC.zip -d .github/ghidra/
+    - name: Install Ghidrathon
+      run : |
+        mkdir ./.github/ghidrathon
+        wget "https://github.com/mandiant/Ghidrathon/releases/download/v${{ matrix.ghidrathon-version }}/Ghidrathon-v${{ matrix.ghidrathon-version}}.zip" -O ./.github/ghidrathon/ghidrathon-v${{ matrix.ghidrathon-version }}.zip
+        unzip .github/ghidrathon/ghidrathon-v${{ matrix.ghidrathon-version }}.zip -d .github/ghidrathon/
+        python -m pip install -r .github/ghidrathon/requirements.txt
+        python .github/ghidrathon/ghidrathon_configure.py $(pwd)/.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC
+        unzip .github/ghidrathon/Ghidrathon-v${{ matrix.ghidrathon-version }}.zip -d .github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC/Ghidra/Extensions
     - name: Install pyyaml
       run: sudo apt-get install -y libyaml-dev
-    - name: Install capa with Ghidra extra
+    - name: Install capa
       run: |
-        pip install -e .[dev,ghidra]
+        pip install -r requirements.txt
+        pip install -e .[dev,scripts]
     - name: Run tests
-      env:
-        GHIDRA_INSTALL_DIR: ${{ github.workspace }}/.github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC
-      run: pytest -v tests/test_ghidra_features.py
+      run: | 
+        mkdir ./.github/ghidra/project
+        .github/ghidra/ghidra_${{ matrix.ghidra-version }}_PUBLIC/support/analyzeHeadless .github/ghidra/project ghidra_test -Import ./tests/data/mimikatz.exe_ -ScriptPath ./tests/ -PostScript test_ghidra_features.py > ../output.log
+        cat ../output.log
+        exit_code=$(cat ../output.log | grep exit | awk '{print $NF}')
+        exit $exit_code
  
+  idalib-tests:
+    name: IDA ${{ matrix.ida.version }} tests for ${{ matrix.python-version }}
+    runs-on: ubuntu-22.04
+    needs: [tests]
+    env:
+      IDA_LICENSE_ID: ${{ secrets.IDA_LICENSE_ID }}
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.10", "3.13"]
+        ida:
+          - version: 9.1
+            slug: "release/9.1/ida-essential/ida-essential_91_x64linux.run"
+          - version: 9.2
+            slug: "release/9.2/ida-essential/ida-essential_92_x64linux.run"
+    steps:
+    - name: Checkout capa with submodules
+      # do only run if IDA_LICENSE_ID is available, have to do this in every step, see https://github.com/orgs/community/discussions/26726#discussioncomment-3253118
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
+      with:
+        submodules: recursive
+    - name: Set up Python ${{ matrix.python-version }}
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      uses: actions/setup-python@0a5c61591373683505ea898e09a3ea4f39ef2b9c # v5.0.0
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Setup uv
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      uses: astral-sh/setup-uv@v6
+    - name: Install dependencies
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      run: sudo apt-get install -y libyaml-dev
+    - name: Install capa
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      run: |
+        pip install -r requirements.txt
+        pip install -e .[dev,scripts]
+        pip install idapro
+    - name: Install IDA ${{ matrix.ida.version }}
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      run: |
+        uv run hcli --disable-updates ida install --download-id ${{ matrix.ida.slug }} --license-id ${{ secrets.IDA_LICENSE_ID }} --set-default --yes
+      env:
+        HCLI_API_KEY: ${{ secrets.HCLI_API_KEY }}
+        IDA_LICENSE_ID: ${{ secrets.IDA_LICENSE_ID }}
+    - name: Run tests
+      if: ${{ env.IDA_LICENSE_ID != 0 }}
+      run: pytest -v tests/test_idalib_features.py  # explicitly refer to the idalib tests for performance. other tests run above.

.pre-commit-config.yaml

@@ -138,6 +138,7 @@ repos:
         -   "--ignore=tests/test_ghidra_features.py"
         -   "--ignore=tests/test_ida_features.py"
         -   "--ignore=tests/test_viv_features.py"
+        -   "--ignore=tests/test_idalib_features.py"
         -   "--ignore=tests/test_main.py"
         -   "--ignore=tests/test_scripts.py"
         always_run: true

CHANGELOG.md

@@ -4,8 +4,6 @@
 
 ### New Features
 
-- ghidra: support PyGhidra @mike-hunhoff #2788
-
 ### Breaking Changes
 
 ### New Rules (4)
@@ -56,6 +54,7 @@ Additionally a Binary Ninja bug has been fixed. Released binaries now include AR
 ### New Features
 
 - ci: add support for arm64 binary releases
+- tests: run tests against IDA via idalib @williballenthin #2742
 
 ### Breaking Changes
 

README.md

@@ -291,17 +291,11 @@ It also uses your local changes to the .idb to extract better features, such as
 ![capa + IDA Pro integration](https://github.com/mandiant/capa/blob/master/doc/img/explorer_expanded.png)
 
 # Ghidra integration
-
-capa supports using Ghidra (via [PyGhidra](https://github.com/NationalSecurityAgency/ghidra/tree/master/Ghidra/Features/PyGhidra)) as a feature extraction backend. This allows you to run capa against binaries using Ghidra's analysis engine.
-
-You can run and view capa results in the Ghidra UI using [capa explorer for Ghidra](https://github.com/mandiant/capa/tree/master/capa/ghidra/plugin).
+If you use Ghidra, then you can use the [capa + Ghidra integration](/capa/ghidra/) to run capa's analysis directly on your Ghidra database and render the results in Ghidra's user interface.
 
 <img src="https://github.com/mandiant/capa/assets/66766340/eeae33f4-99d4-42dc-a5e8-4c1b8c661492" width=300>
 
-You can also run capa from the command line using the [Ghidra backend](https://github.com/mandiant/capa/tree/master/capa/ghidra).
-
 # blog posts
-- [Riding Dragons: capa Harnesses Ghidra](https://www.mandiant.com/resources/blog/capa-harnesses-ghidra)
 - [Dynamic capa: Exploring Executable Run-Time Behavior with the CAPE Sandbox](https://www.mandiant.com/resources/blog/dynamic-capa-executable-behavior-cape-sandbox)
 - [capa v4: casting a wider .NET](https://www.mandiant.com/resources/blog/capa-v4-casting-wider-net) (.NET support)
 - [ELFant in the Room – capa v3](https://www.mandiant.com/resources/elfant-in-the-room-capa-v3) (ELF support)

capa/features/extractors/ghidra/basicblock.py

@@ -83,7 +83,7 @@ def bb_contains_stackstring(bb: ghidra.program.model.block.CodeBlock) -> bool:
     true if basic block contains enough moves of constant bytes to the stack
     """
     count = 0
-    for insn in capa.features.extractors.ghidra.helpers.get_current_program().getListing().getInstructions(bb, True):
+    for insn in currentProgram().getListing().getInstructions(bb, True):  # type: ignore [name-defined] # noqa: F821
         if is_mov_imm_to_stack(insn):
             count += get_printable_len(insn.getScalar(1))
         if count > MIN_STACKSTRING_LEN:
@@ -96,9 +96,7 @@ def _bb_has_tight_loop(bb: ghidra.program.model.block.CodeBlock):
     parse tight loops, true if last instruction in basic block branches to bb start
     """
     # Reverse Ordered, first InstructionDB
-    last_insn = (
-        capa.features.extractors.ghidra.helpers.get_current_program().getListing().getInstructions(bb, False).next()
-    )
+    last_insn = currentProgram().getListing().getInstructions(bb, False).next()  # type: ignore [name-defined] # noqa: F821
 
     if last_insn.getFlowType().isJump():
         return last_insn.getAddress(0) == bb.getMinAddress()
@@ -142,3 +140,20 @@ def extract_features(fh: FunctionHandle, bbh: BBHandle) -> Iterator[tuple[Featur
     for bb_handler in BASIC_BLOCK_HANDLERS:
         for feature, addr in bb_handler(fh, bbh):
             yield feature, addr
+
+
+def main():
+    features = []
+    from capa.features.extractors.ghidra.extractor import GhidraFeatureExtractor
+
+    for fh in GhidraFeatureExtractor().get_functions():
+        for bbh in capa.features.extractors.ghidra.helpers.get_function_blocks(fh):
+            features.extend(list(extract_features(fh, bbh)))
+
+    import pprint
+
+    pprint.pprint(features)  # noqa: T203
+
+
+if __name__ == "__main__":
+    main()

capa/features/extractors/ghidra/context.py

@@ -1,44 +0,0 @@
-# Copyright 2023 Google LLC
-#
-# 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
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# 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
-
-
-class GhidraContext:
-    """
-    State holder for the Ghidra backend to avoid passing state to every function.
-
-    PyGhidra uses a context manager to set up the Ghidra environment (program, transaction, etc.).
-    We store the relevant objects here to allow easy access throughout the extractor
-    without needing to pass them as arguments to every feature extraction method.
-    """
-
-    def __init__(self, program, flat_api, monitor):
-        self.program = program
-        self.flat_api = flat_api
-        self.monitor = monitor
-
-
-_context: Optional[GhidraContext] = None
-
-
-def set_context(program, flat_api, monitor):
-    global _context
-    _context = GhidraContext(program, flat_api, monitor)
-
-
-def get_context() -> GhidraContext:
-    if _context is None:
-        raise RuntimeError("GhidraContext not initialized")
-    return _context

capa/features/extractors/ghidra/extractor.py

@@ -12,14 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import weakref
-import contextlib
 from typing import Iterator
 
 import capa.features.extractors.ghidra.file
 import capa.features.extractors.ghidra.insn
 import capa.features.extractors.ghidra.global_
-import capa.features.extractors.ghidra.helpers as ghidra_helpers
 import capa.features.extractors.ghidra.function
 import capa.features.extractors.ghidra.basicblock
 from capa.features.common import Feature
@@ -34,20 +31,19 @@ from capa.features.extractors.base_extractor import (
 
 
 class GhidraFeatureExtractor(StaticFeatureExtractor):
-    def __init__(self, ctx_manager=None, tmpdir=None):
-        self.ctx_manager = ctx_manager
-        self.tmpdir = tmpdir
+    def __init__(self):
+        import capa.features.extractors.ghidra.helpers as ghidra_helpers
 
         super().__init__(
             SampleHashes(
-                md5=ghidra_helpers.get_current_program().getExecutableMD5(),
+                md5=capa.ghidra.helpers.get_file_md5(),
                 # ghidra doesn't expose this hash.
                 # https://ghidra.re/ghidra_docs/api/ghidra/program/model/listing/Program.html
                 #
                 # the hashes are stored in the database, not computed on the fly,
                 # so it's probably not trivial to add SHA1.
                 sha1="",
-                sha256=ghidra_helpers.get_current_program().getExecutableSHA256(),
+                sha256=capa.ghidra.helpers.get_file_sha256(),
             )
         )
 
@@ -59,14 +55,8 @@ class GhidraFeatureExtractor(StaticFeatureExtractor):
         self.externs = ghidra_helpers.get_file_externs()
         self.fakes = ghidra_helpers.map_fake_import_addrs()
 
-        # Register cleanup to run when the extractor is garbage collected or when the program exits.
-        # We use weakref.finalize instead of __del__ to avoid issues with reference cycles and
-        # to ensure deterministic cleanup on interpreter shutdown.
-        if self.ctx_manager or self.tmpdir:
-            weakref.finalize(self, cleanup, self.ctx_manager, self.tmpdir)
-
     def get_base_address(self):
-        return AbsoluteVirtualAddress(ghidra_helpers.get_current_program().getImageBase().getOffset())
+        return AbsoluteVirtualAddress(currentProgram().getImageBase().getOffset())  # type: ignore [name-defined] # noqa: F821
 
     def extract_global_features(self):
         yield from self.global_features
@@ -75,6 +65,7 @@ class GhidraFeatureExtractor(StaticFeatureExtractor):
         yield from capa.features.extractors.ghidra.file.extract_features()
 
     def get_functions(self) -> Iterator[FunctionHandle]:
+        import capa.features.extractors.ghidra.helpers as ghidra_helpers
 
         for fhandle in ghidra_helpers.get_function_symbols():
             fh: FunctionHandle = FunctionHandle(
@@ -86,14 +77,14 @@ class GhidraFeatureExtractor(StaticFeatureExtractor):
 
     @staticmethod
     def get_function(addr: int) -> FunctionHandle:
-
-        func = ghidra_helpers.get_flat_api().getFunctionContaining(ghidra_helpers.get_flat_api().toAddr(addr))
+        func = getFunctionContaining(toAddr(addr))  # type: ignore [name-defined] # noqa: F821
         return FunctionHandle(address=AbsoluteVirtualAddress(func.getEntryPoint().getOffset()), inner=func)
 
     def extract_function_features(self, fh: FunctionHandle) -> Iterator[tuple[Feature, Address]]:
         yield from capa.features.extractors.ghidra.function.extract_features(fh)
 
     def get_basic_blocks(self, fh: FunctionHandle) -> Iterator[BBHandle]:
+        import capa.features.extractors.ghidra.helpers as ghidra_helpers
 
         yield from ghidra_helpers.get_function_blocks(fh)
 
@@ -101,17 +92,9 @@ class GhidraFeatureExtractor(StaticFeatureExtractor):
         yield from capa.features.extractors.ghidra.basicblock.extract_features(fh, bbh)
 
     def get_instructions(self, fh: FunctionHandle, bbh: BBHandle) -> Iterator[InsnHandle]:
+        import capa.features.extractors.ghidra.helpers as ghidra_helpers
 
         yield from ghidra_helpers.get_insn_in_range(bbh)
 
     def extract_insn_features(self, fh: FunctionHandle, bbh: BBHandle, ih: InsnHandle):
         yield from capa.features.extractors.ghidra.insn.extract_features(fh, bbh, ih)
-
-
-def cleanup(ctx_manager, tmpdir):
-    if ctx_manager:
-        with contextlib.suppress(Exception):
-            ctx_manager.__exit__(None, None, None)
-    if tmpdir:
-        with contextlib.suppress(Exception):
-            tmpdir.cleanup()

capa/features/extractors/ghidra/file.py

@@ -80,54 +80,22 @@ def extract_file_embedded_pe() -> Iterator[tuple[Feature, Address]]:
         for i in range(256)
     ]
 
-    for block in capa.features.extractors.ghidra.helpers.get_current_program().getMemory().getBlocks():
+    for block in currentProgram().getMemory().getBlocks():  # type: ignore [name-defined] # noqa: F821
         if not all((block.isLoaded(), block.isInitialized(), "Headers" not in block.getName())):
             continue
 
         for off, _ in find_embedded_pe(capa.features.extractors.ghidra.helpers.get_block_bytes(block), mz_xor):
             # add offset back to block start
-            ea_addr = block.getStart().add(off)
-            ea = ea_addr.getOffset()
-            f_offset = capa.features.extractors.ghidra.helpers.get_file_offset(ea_addr)
-            if f_offset != -1:
-                ea = f_offset
+            ea: int = block.getStart().add(off).getOffset()
 
             yield Characteristic("embedded pe"), FileOffsetAddress(ea)
 
 
 def extract_file_export_names() -> Iterator[tuple[Feature, Address]]:
     """extract function exports"""
-    program = capa.features.extractors.ghidra.helpers.get_current_program()
-    st = program.getSymbolTable()
-
+    st = currentProgram().getSymbolTable()  # type: ignore [name-defined] # noqa: F821
     for addr in st.getExternalEntryPointIterator():
-        sym = st.getPrimarySymbol(addr)
-        name = sym.getName()
-
-        # Check for forwarded export
-        is_forwarded = False
-        refs = program.getReferenceManager().getReferencesFrom(addr)
-        for ref in refs:
-            if ref.getToAddress().isExternalAddress():
-                ext_sym = st.getPrimarySymbol(ref.getToAddress())
-                if ext_sym:
-                    ext_loc = program.getExternalManager().getExternalLocation(ext_sym)
-                    if ext_loc:
-                        # It is a forwarded export
-                        libname = ext_loc.getLibraryName()
-                        if libname.lower().endswith(".dll"):
-                            libname = libname[:-4]
-
-                        forwarded_name = f"{libname}.{ext_loc.getLabel()}"
-                        forwarded_name = capa.features.extractors.helpers.reformat_forwarded_export_name(forwarded_name)
-
-                        yield Export(forwarded_name), AbsoluteVirtualAddress(addr.getOffset())
-                        yield Characteristic("forwarded export"), AbsoluteVirtualAddress(addr.getOffset())
-                        is_forwarded = True
-                        break
-
-        if not is_forwarded:
-            yield Export(name), AbsoluteVirtualAddress(addr.getOffset())
+        yield Export(st.getPrimarySymbol(addr).getName()), AbsoluteVirtualAddress(addr.getOffset())
 
 
 def extract_file_import_names() -> Iterator[tuple[Feature, Address]]:
@@ -142,7 +110,7 @@ def extract_file_import_names() -> Iterator[tuple[Feature, Address]]:
      - importname
     """
 
-    for f in capa.features.extractors.ghidra.helpers.get_current_program().getFunctionManager().getExternalFunctions():
+    for f in currentProgram().getFunctionManager().getExternalFunctions():  # type: ignore [name-defined] # noqa: F821
         for r in f.getSymbol().getReferences():
             if r.getReferenceType().isData():
                 addr = r.getFromAddress().getOffset()  # gets pointer to fake external addr
@@ -158,14 +126,14 @@ def extract_file_import_names() -> Iterator[tuple[Feature, Address]]:
 def extract_file_section_names() -> Iterator[tuple[Feature, Address]]:
     """extract section names"""
 
-    for block in capa.features.extractors.ghidra.helpers.get_current_program().getMemory().getBlocks():
+    for block in currentProgram().getMemory().getBlocks():  # type: ignore [name-defined] # noqa: F821
         yield Section(block.getName()), AbsoluteVirtualAddress(block.getStart().getOffset())
 
 
 def extract_file_strings() -> Iterator[tuple[Feature, Address]]:
     """extract ASCII and UTF-16 LE strings"""
 
-    for block in capa.features.extractors.ghidra.helpers.get_current_program().getMemory().getBlocks():
+    for block in currentProgram().getMemory().getBlocks():  # type: ignore [name-defined] # noqa: F821
         if not block.isInitialized():
             continue
 
@@ -185,8 +153,7 @@ def extract_file_function_names() -> Iterator[tuple[Feature, Address]]:
     extract the names of statically-linked library functions.
     """
 
-    for sym in capa.features.extractors.ghidra.helpers.get_current_program().getSymbolTable().getAllSymbols(True):
-
+    for sym in currentProgram().getSymbolTable().getAllSymbols(True):  # type: ignore [name-defined] # noqa: F821
         # .isExternal() misses more than this config for the function symbols
         if sym.getSymbolType() == SymbolType.FUNCTION and sym.getSource() == SourceType.ANALYSIS and sym.isGlobal():
             name = sym.getName()  # starts to resolve names based on Ghidra's FidDB
@@ -203,7 +170,7 @@ def extract_file_function_names() -> Iterator[tuple[Feature, Address]]:
 
 
 def extract_file_format() -> Iterator[tuple[Feature, Address]]:
-    ef = capa.features.extractors.ghidra.helpers.get_current_program().getExecutableFormat()
+    ef = currentProgram().getExecutableFormat()  # type: ignore [name-defined] # noqa: F821
     if "PE" in ef:
         yield Format(FORMAT_PE), NO_ADDRESS
     elif "ELF" in ef:
@@ -231,3 +198,14 @@ FILE_HANDLERS = (
     extract_file_function_names,
     extract_file_format,
 )
+
+
+def main():
+    """ """
+    import pprint
+
+    pprint.pprint(list(extract_features()))  # noqa: T203
+
+
+if __name__ == "__main__":
+    main()

capa/features/extractors/ghidra/function.py

@@ -26,25 +26,21 @@ from capa.features.extractors.base_extractor import FunctionHandle
 
 def extract_function_calls_to(fh: FunctionHandle):
     """extract callers to a function"""
-    f: "ghidra.program.database.function.FunctionDB" = fh.inner
+    f: ghidra.program.database.function.FunctionDB = fh.inner
     for ref in f.getSymbol().getReferences():
         if ref.getReferenceType().isCall():
             yield Characteristic("calls to"), AbsoluteVirtualAddress(ref.getFromAddress().getOffset())
 
 
 def extract_function_loop(fh: FunctionHandle):
-    f: "ghidra.program.database.function.FunctionDB" = fh.inner
+    f: ghidra.program.database.function.FunctionDB = fh.inner
 
     edges = []
-    for block in SimpleBlockIterator(
-        BasicBlockModel(capa.features.extractors.ghidra.helpers.get_current_program()),
-        f.getBody(),
-        capa.features.extractors.ghidra.helpers.get_monitor(),
-    ):
-        dests = block.getDestinations(capa.features.extractors.ghidra.helpers.get_monitor())
+    for block in SimpleBlockIterator(BasicBlockModel(currentProgram()), f.getBody(), monitor()):  # type: ignore [name-defined] # noqa: F821
+        dests = block.getDestinations(monitor())  # type: ignore [name-defined] # noqa: F821
         s_addrs = block.getStartAddresses()
 
-        while dests.hasNext():
+        while dests.hasNext():  # For loop throws Python TypeError
             for addr in s_addrs:
                 edges.append((addr.getOffset(), dests.next().getDestinationAddress().getOffset()))
 
@@ -53,17 +49,32 @@ def extract_function_loop(fh: FunctionHandle):
 
 
 def extract_recursive_call(fh: FunctionHandle):
-    f: "ghidra.program.database.function.FunctionDB" = fh.inner
+    f: ghidra.program.database.function.FunctionDB = fh.inner
 
-    for func in f.getCalledFunctions(capa.features.extractors.ghidra.helpers.get_monitor()):
+    for func in f.getCalledFunctions(monitor()):  # type: ignore [name-defined] # noqa: F821
         if func.getEntryPoint().getOffset() == f.getEntryPoint().getOffset():
             yield Characteristic("recursive call"), AbsoluteVirtualAddress(f.getEntryPoint().getOffset())
 
 
 def extract_features(fh: FunctionHandle) -> Iterator[tuple[Feature, Address]]:
-    for function_handler in FUNCTION_HANDLERS:
-        for feature, addr in function_handler(fh):
+    for func_handler in FUNCTION_HANDLERS:
+        for feature, addr in func_handler(fh):
             yield feature, addr
 
 
 FUNCTION_HANDLERS = (extract_function_calls_to, extract_function_loop, extract_recursive_call)
+
+
+def main():
+    """ """
+    features = []
+    for fhandle in capa.features.extractors.ghidra.helpers.get_function_symbols():
+        features.extend(list(extract_features(fhandle)))
+
+    import pprint
+
+    pprint.pprint(features)  # noqa: T203
+
+
+if __name__ == "__main__":
+    main()

capa/features/extractors/ghidra/global_.py

@@ -26,7 +26,7 @@ logger = logging.getLogger(__name__)
 
 
 def extract_os() -> Iterator[tuple[Feature, Address]]:
-    format_name: str = capa.features.extractors.ghidra.helpers.get_current_program().getExecutableFormat()
+    format_name: str = currentProgram().getExecutableFormat()  # type: ignore [name-defined] # noqa: F821
 
     if "PE" in format_name:
         yield OS(OS_WINDOWS), NO_ADDRESS
@@ -53,7 +53,7 @@ def extract_os() -> Iterator[tuple[Feature, Address]]:
 
 
 def extract_arch() -> Iterator[tuple[Feature, Address]]:
-    lang_id = capa.features.extractors.ghidra.helpers.get_current_program().getMetadata().get("Language ID")
+    lang_id = currentProgram().getMetadata().get("Language ID")  # type: ignore [name-defined] # noqa: F821
 
     if "x86" in lang_id and "64" in lang_id:
         yield Arch(ARCH_AMD64), NO_ADDRESS

capa/features/extractors/ghidra/helpers.py

@@ -22,22 +22,9 @@ from ghidra.program.model.symbol import SourceType, SymbolType
 from ghidra.program.model.address import AddressSpace
 
 import capa.features.extractors.helpers
-import capa.features.extractors.ghidra.context as ghidra_context
 from capa.features.common import THUNK_CHAIN_DEPTH_DELTA
 from capa.features.address import AbsoluteVirtualAddress
-from capa.features.extractors.base_extractor import BBHandle, InsnHandle
-
-
-def get_current_program():
-    return ghidra_context.get_context().program
-
-
-def get_monitor():
-    return ghidra_context.get_context().monitor
-
-
-def get_flat_api():
-    return ghidra_context.get_context().flat_api
+from capa.features.extractors.base_extractor import BBHandle, InsnHandle, FunctionHandle
 
 
 def ints_to_bytes(bytez: list[int]) -> bytes:
@@ -49,7 +36,7 @@ def ints_to_bytes(bytez: list[int]) -> bytes:
     return bytes([b & 0xFF for b in bytez])
 
 
-def find_byte_sequence(addr: "ghidra.program.model.address.Address", seq: bytes) -> Iterator[int]:
+def find_byte_sequence(addr: ghidra.program.model.address.Address, seq: bytes) -> Iterator[int]:
     """yield all ea of a given byte sequence
 
     args:
@@ -57,25 +44,12 @@ def find_byte_sequence(addr: "ghidra.program.model.address.Address", seq: bytes)
         seq: bytes to search e.g. b"\x01\x03"
     """
     seqstr = "".join([f"\\x{b:02x}" for b in seq])
-    eas = get_flat_api().findBytes(addr, seqstr, java.lang.Integer.MAX_VALUE, 1)
+    eas = findBytes(addr, seqstr, java.lang.Integer.MAX_VALUE, 1)  # type: ignore [name-defined] # noqa: F821
 
     yield from eas
 
 
-def get_file_offset(addr: "ghidra.program.model.address.Address") -> int:
-    """get file offset for an address"""
-    block = get_current_program().getMemory().getBlock(addr)
-    if not block:
-        return -1
-
-    for info in block.getSourceInfos():
-        if info.contains(addr):
-            return info.getFileBytesOffset(addr)
-
-    return -1
-
-
-def get_bytes(addr: "ghidra.program.model.address.Address", length: int) -> bytes:
+def get_bytes(addr: ghidra.program.model.address.Address, length: int) -> bytes:
     """yield length bytes at addr
 
     args:
@@ -83,12 +57,12 @@ def get_bytes(addr: "ghidra.program.model.address.Address", length: int) -> byte
         length: length of bytes to pull
     """
     try:
-        return ints_to_bytes(get_flat_api().getBytes(addr, int(length)))
-    except Exception:
+        return ints_to_bytes(getBytes(addr, length))  # type: ignore [name-defined] # noqa: F821
+    except RuntimeError:
         return b""
 
 
-def get_block_bytes(block: "ghidra.program.model.mem.MemoryBlock") -> bytes:
+def get_block_bytes(block: ghidra.program.model.mem.MemoryBlock) -> bytes:
     """yield all bytes in a given block
 
     args:
@@ -99,21 +73,20 @@ def get_block_bytes(block: "ghidra.program.model.mem.MemoryBlock") -> bytes:
 
 def get_function_symbols():
     """yield all non-external function symbols"""
-    yield from get_current_program().getFunctionManager().getFunctionsNoStubs(True)
+    yield from currentProgram().getFunctionManager().getFunctionsNoStubs(True)  # type: ignore [name-defined] # noqa: F821
 
 
-def get_function_blocks(fh: "capa.features.extractors.base_extractor.FunctionHandle") -> Iterator[BBHandle]:
-    """
-    yield the basic blocks of the function
-    """
+def get_function_blocks(fh: FunctionHandle) -> Iterator[BBHandle]:
+    """yield BBHandle for each bb in a given function"""
 
-    for block in SimpleBlockIterator(BasicBlockModel(get_current_program()), fh.inner.getBody(), get_monitor()):
-        yield BBHandle(address=AbsoluteVirtualAddress(block.getMinAddress().getOffset()), inner=block)
+    func: ghidra.program.database.function.FunctionDB = fh.inner
+    for bb in SimpleBlockIterator(BasicBlockModel(currentProgram()), func.getBody(), monitor()):  # type: ignore [name-defined] # noqa: F821
+        yield BBHandle(address=AbsoluteVirtualAddress(bb.getMinAddress().getOffset()), inner=bb)
 
 
 def get_insn_in_range(bbh: BBHandle) -> Iterator[InsnHandle]:
     """yield InshHandle for each insn in a given basicblock"""
-    for insn in get_current_program().getListing().getInstructions(bbh.inner, True):
+    for insn in currentProgram().getListing().getInstructions(bbh.inner, True):  # type: ignore [name-defined] # noqa: F821
         yield InsnHandle(address=AbsoluteVirtualAddress(insn.getAddress().getOffset()), inner=insn)
 
 
@@ -122,7 +95,7 @@ def get_file_imports() -> dict[int, list[str]]:
 
     import_dict: dict[int, list[str]] = {}
 
-    for f in get_current_program().getFunctionManager().getExternalFunctions():
+    for f in currentProgram().getFunctionManager().getExternalFunctions():  # type: ignore [name-defined] # noqa: F821
         for r in f.getSymbol().getReferences():
             if r.getReferenceType().isData():
                 addr = r.getFromAddress().getOffset()  # gets pointer to fake external addr
@@ -160,7 +133,7 @@ def get_file_externs() -> dict[int, list[str]]:
 
     extern_dict: dict[int, list[str]] = {}
 
-    for sym in get_current_program().getSymbolTable().getAllSymbols(True):
+    for sym in currentProgram().getSymbolTable().getAllSymbols(True):  # type: ignore [name-defined] # noqa: F821
         # .isExternal() misses more than this config for the function symbols
         if sym.getSymbolType() == SymbolType.FUNCTION and sym.getSource() == SourceType.ANALYSIS and sym.isGlobal():
             name = sym.getName()  # starts to resolve names based on Ghidra's FidDB
@@ -198,7 +171,7 @@ def map_fake_import_addrs() -> dict[int, list[int]]:
     """
     fake_dict: dict[int, list[int]] = {}
 
-    for f in get_current_program().getFunctionManager().getExternalFunctions():
+    for f in currentProgram().getFunctionManager().getExternalFunctions():  # type: ignore [name-defined] # noqa: F821
         for r in f.getSymbol().getReferences():
             if r.getReferenceType().isData():
                 fake_dict.setdefault(f.getEntryPoint().getOffset(), []).append(r.getFromAddress().getOffset())
@@ -207,7 +180,7 @@ def map_fake_import_addrs() -> dict[int, list[int]]:
 
 
 def check_addr_for_api(
-    addr: "ghidra.program.model.address.Address",
+    addr: ghidra.program.model.address.Address,
     fakes: dict[int, list[int]],
     imports: dict[int, list[str]],
     externs: dict[int, list[str]],
@@ -229,18 +202,18 @@ def check_addr_for_api(
     return False
 
 
-def is_call_or_jmp(insn: "ghidra.program.database.code.InstructionDB") -> bool:
+def is_call_or_jmp(insn: ghidra.program.database.code.InstructionDB) -> bool:
     return any(mnem in insn.getMnemonicString() for mnem in ["CALL", "J"])  # JMP, JNE, JNZ, etc
 
 
-def is_sp_modified(insn: "ghidra.program.database.code.InstructionDB") -> bool:
+def is_sp_modified(insn: ghidra.program.database.code.InstructionDB) -> bool:
     for i in range(insn.getNumOperands()):
         if insn.getOperandType(i) == OperandType.REGISTER:
             return "SP" in insn.getRegister(i).getName() and insn.getOperandRefType(i).isWrite()
     return False
 
 
-def is_stack_referenced(insn: "ghidra.program.database.code.InstructionDB") -> bool:
+def is_stack_referenced(insn: ghidra.program.database.code.InstructionDB) -> bool:
     """generic catch-all for stack references"""
     for i in range(insn.getNumOperands()):
         if insn.getOperandType(i) == OperandType.REGISTER:
@@ -252,7 +225,7 @@ def is_stack_referenced(insn: "ghidra.program.database.code.InstructionDB") -> b
     return any(ref.isStackReference() for ref in insn.getReferencesFrom())
 
 
-def is_zxor(insn: "ghidra.program.database.code.InstructionDB") -> bool:
+def is_zxor(insn: ghidra.program.database.code.InstructionDB) -> bool:
     # assume XOR insn
     # XOR's against the same operand zero out
     ops = []
@@ -268,29 +241,29 @@ def is_zxor(insn: "ghidra.program.database.code.InstructionDB") -> bool:
     return all(n == operands[0] for n in operands)
 
 
-def handle_thunk(addr: "ghidra.program.model.address.Address"):
+def handle_thunk(addr: ghidra.program.model.address.Address):
     """Follow thunk chains down to a reasonable depth"""
     ref = addr
     for _ in range(THUNK_CHAIN_DEPTH_DELTA):
-        thunk_jmp = get_flat_api().getInstructionAt(ref)
+        thunk_jmp = getInstructionAt(ref)  # type: ignore [name-defined] # noqa: F821
         if thunk_jmp and is_call_or_jmp(thunk_jmp):
             if OperandType.isAddress(thunk_jmp.getOperandType(0)):
                 ref = thunk_jmp.getAddress(0)
         else:
-            thunk_dat = get_flat_api().getDataContaining(ref)
+            thunk_dat = getDataContaining(ref)  # type: ignore [name-defined] # noqa: F821
             if thunk_dat and thunk_dat.isDefined() and thunk_dat.isPointer():
                 ref = thunk_dat.getValue()
                 break  # end of thunk chain reached
     return ref
 
 
-def dereference_ptr(insn: "ghidra.program.database.code.InstructionDB"):
+def dereference_ptr(insn: ghidra.program.database.code.InstructionDB):
     addr_code = OperandType.ADDRESS | OperandType.CODE
     to_deref = insn.getAddress(0)
-    dat = get_flat_api().getDataContaining(to_deref)
+    dat = getDataContaining(to_deref)  # type: ignore [name-defined] # noqa: F821
 
     if insn.getOperandType(0) == addr_code:
-        thfunc = get_flat_api().getFunctionContaining(to_deref)
+        thfunc = getFunctionContaining(to_deref)  # type: ignore [name-defined] # noqa: F821
         if thfunc and thfunc.isThunk():
             return handle_thunk(to_deref)
         else:
@@ -321,7 +294,7 @@ def find_data_references_from_insn(insn, max_depth: int = 10):
         to_addr = reference.getToAddress()
 
         for _ in range(max_depth - 1):
-            data = get_flat_api().getDataAt(to_addr)
+            data = getDataAt(to_addr)  # type: ignore [name-defined] # noqa: F821
             if data and data.isPointer():
                 ptr_value = data.getValue()
 

capa/features/extractors/ghidra/insn.py

@@ -234,7 +234,7 @@ def extract_insn_bytes_features(fh: FunctionHandle, bb: BBHandle, ih: InsnHandle
         push    offset iid_004118d4_IShellLinkA ; riid
     """
     for addr in capa.features.extractors.ghidra.helpers.find_data_references_from_insn(ih.inner):
-        data = capa.features.extractors.ghidra.helpers.get_flat_api().getDataAt(addr)
+        data = getDataAt(addr)  # type: ignore [name-defined] # noqa: F821
         if data and not data.hasStringValue():
             extracted_bytes = capa.features.extractors.ghidra.helpers.get_bytes(addr, MAX_BYTES_FEATURE_SIZE)
             if extracted_bytes and not capa.features.extractors.helpers.all_zeros(extracted_bytes):
@@ -249,9 +249,9 @@ def extract_insn_string_features(fh: FunctionHandle, bb: BBHandle, ih: InsnHandl
         push offset aAcr     ; "ACR  > "
     """
     for addr in capa.features.extractors.ghidra.helpers.find_data_references_from_insn(ih.inner):
-        data = capa.features.extractors.ghidra.helpers.get_flat_api().getDataAt(addr)
+        data = getDataAt(addr)  # type: ignore [name-defined] # noqa: F821
         if data and data.hasStringValue():
-            yield String(str(data.getValue())), ih.address
+            yield String(data.getValue()), ih.address
 
 
 def extract_insn_mnemonic_features(
@@ -361,8 +361,8 @@ def extract_insn_cross_section_cflow(
         if capa.features.extractors.ghidra.helpers.check_addr_for_api(ref, fakes, imports, externs):
             return
 
-    this_mem_block = capa.features.extractors.ghidra.helpers.get_flat_api().getMemoryBlock(insn.getAddress())
-    ref_block = capa.features.extractors.ghidra.helpers.get_flat_api().getMemoryBlock(ref)
+    this_mem_block = getMemoryBlock(insn.getAddress())  # type: ignore [name-defined] # noqa: F821
+    ref_block = getMemoryBlock(ref)  # type: ignore [name-defined] # noqa: F821
     if ref_block != this_mem_block:
         yield Characteristic("cross section flow"), ih.address
 
@@ -425,19 +425,19 @@ def check_nzxor_security_cookie_delta(
     Check if insn within last addr of last bb - delta
     """
 
-    model = SimpleBlockModel(capa.features.extractors.ghidra.helpers.get_current_program())
+    model = SimpleBlockModel(currentProgram())  # type: ignore [name-defined] # noqa: F821
     insn_addr = insn.getAddress()
     func_asv = fh.getBody()
 
     first_addr = func_asv.getMinAddress()
     if insn_addr < first_addr.add(SECURITY_COOKIE_BYTES_DELTA):
-        first_bb = model.getFirstCodeBlockContaining(first_addr, capa.features.extractors.ghidra.helpers.get_monitor())
+        first_bb = model.getFirstCodeBlockContaining(first_addr, monitor())  # type: ignore [name-defined] # noqa: F821
         if first_bb.contains(insn_addr):
             return True
 
     last_addr = func_asv.getMaxAddress()
     if insn_addr > last_addr.add(SECURITY_COOKIE_BYTES_DELTA * -1):
-        last_bb = model.getFirstCodeBlockContaining(last_addr, capa.features.extractors.ghidra.helpers.get_monitor())
+        last_bb = model.getFirstCodeBlockContaining(last_addr, monitor())  # type: ignore [name-defined] # noqa: F821
         if last_bb.contains(insn_addr):
             return True
 
@@ -488,3 +488,22 @@ INSTRUCTION_HANDLERS = (
     extract_function_calls_from,
     extract_function_indirect_call_characteristic_features,
 )
+
+
+def main():
+    """ """
+    features = []
+    from capa.features.extractors.ghidra.extractor import GhidraFeatureExtractor
+
+    for fh in GhidraFeatureExtractor().get_functions():
+        for bb in capa.features.extractors.ghidra.helpers.get_function_blocks(fh):
+            for insn in capa.features.extractors.ghidra.helpers.get_insn_in_range(bb):
+                features.extend(list(extract_features(fh, bb, insn)))
+
+    import pprint
+
+    pprint.pprint(features)  # noqa: T203
+
+
+if __name__ == "__main__":
+    main()

capa/features/extractors/ida/function.py

@@ -18,6 +18,7 @@ import idaapi
 import idautils
 
 import capa.features.extractors.ida.helpers
+from capa.features.file import FunctionName
 from capa.features.common import Feature, Characteristic
 from capa.features.address import Address, AbsoluteVirtualAddress
 from capa.features.extractors import loops
@@ -50,10 +51,35 @@ def extract_recursive_call(fh: FunctionHandle):
         yield Characteristic("recursive call"), fh.address
 
 
+def extract_function_name(fh: FunctionHandle) -> Iterator[tuple[Feature, Address]]:
+    ea = fh.inner.start_ea
+    name = idaapi.get_name(ea)
+    yield FunctionName(name), fh.address
+    if name.startswith("_"):
+        # some linkers may prefix linked routines with a `_` to avoid name collisions.
+        # extract features for both the mangled and un-mangled representations.
+        # e.g. `_fwrite` -> `fwrite`
+        # see: https://stackoverflow.com/a/2628384/87207
+        yield FunctionName(name[1:]), fh.address
+
+
+def extract_function_alternative_names(fh: FunctionHandle):
+    """Get all alternative names for an address."""
+
+    for aname in capa.features.extractors.ida.helpers.get_function_alternative_names(fh.inner.start_ea):
+        yield FunctionName(aname), fh.address
+
+
 def extract_features(fh: FunctionHandle) -> Iterator[tuple[Feature, Address]]:
     for func_handler in FUNCTION_HANDLERS:
         for feature, addr in func_handler(fh):
             yield feature, addr
 
 
-FUNCTION_HANDLERS = (extract_function_calls_to, extract_function_loop, extract_recursive_call)
+FUNCTION_HANDLERS = (
+    extract_function_calls_to,
+    extract_function_loop,
+    extract_recursive_call,
+    extract_function_name,
+    extract_function_alternative_names,
+)

capa/features/extractors/ida/helpers.py

@@ -20,6 +20,7 @@ import idaapi
 import ida_nalt
 import idautils
 import ida_bytes
+import ida_funcs
 import ida_segment
 
 from capa.features.address import AbsoluteVirtualAddress
@@ -436,3 +437,23 @@ def is_basic_block_return(bb: idaapi.BasicBlock) -> bool:
 def has_sib(oper: idaapi.op_t) -> bool:
     # via: https://reverseengineering.stackexchange.com/a/14300
     return oper.specflag1 == 1
+
+
+def get_function_alternative_names(fva: int):
+    """Get all alternative names for an address."""
+
+    # Check indented comment
+    cmt = ida_bytes.get_cmt(fva, False)  # False = non-repeatable
+    if cmt:
+        for line in cmt.split("\n"):
+            if line.startswith("Alternative name is '") and line.endswith("'"):
+                name = line[len("Alternative name is '") : -1]  # Extract name between quotes
+                yield name
+
+    # Check function comment
+    func_cmt = ida_funcs.get_func_cmt(idaapi.get_func(fva), False)
+    if func_cmt:
+        for line in func_cmt.split("\n"):
+            if line.startswith("Alternative name is '") and line.endswith("'"):
+                name = line[len("Alternative name is '") : -1]
+                yield name

capa/features/extractors/ida/insn.py

@@ -22,6 +22,7 @@ import idautils
 
 import capa.features.extractors.helpers
 import capa.features.extractors.ida.helpers
+from capa.features.file import FunctionName
 from capa.features.insn import API, MAX_STRUCTURE_SIZE, Number, Offset, Mnemonic, OperandNumber, OperandOffset
 from capa.features.common import MAX_BYTES_FEATURE_SIZE, THUNK_CHAIN_DEPTH_DELTA, Bytes, String, Feature, Characteristic
 from capa.features.address import Address, AbsoluteVirtualAddress
@@ -129,8 +130,8 @@ def extract_insn_api_features(fh: FunctionHandle, bbh: BBHandle, ih: InsnHandle)
         # not a function (start)
         return
 
-    if target_func.flags & idaapi.FUNC_LIB:
-        name = idaapi.get_name(target_func.start_ea)
+    name = idaapi.get_name(target_func.start_ea)
+    if target_func.flags & idaapi.FUNC_LIB or not name.startswith("sub_"):
         yield API(name), ih.address
         if name.startswith("_"):
             # some linkers may prefix linked routines with a `_` to avoid name collisions.
@@ -139,6 +140,10 @@ def extract_insn_api_features(fh: FunctionHandle, bbh: BBHandle, ih: InsnHandle)
             # see: https://stackoverflow.com/a/2628384/87207
             yield API(name[1:]), ih.address
 
+        for altname in capa.features.extractors.ida.helpers.get_function_alternative_names(target_func.start_ea):
+            yield FunctionName(altname), ih.address
+            yield API(altname), ih.address
+
 
 def extract_insn_number_features(
     fh: FunctionHandle, bbh: BBHandle, ih: InsnHandle

capa/ghidra/README.md

@@ -1,75 +1,107 @@
-# capa analysis using Ghidra
+<div align="center">
+    <img src="../../doc/img/ghidra_backend_logo.png" width=240 height=125>
+</div>
 
-capa supports using Ghidra (via [PyGhidra](https://github.com/NationalSecurityAgency/ghidra/tree/master/Ghidra/Features/PyGhidra)) as a feature extraction backend. This enables you to run capa against binaries using Ghidra's analysis engine.
+# capa + Ghidra
 
+[capa](https://github.com/mandiant/capa) is the FLARE team’s open-source tool that detects capabilities in executable files. [Ghidra](https://github.com/NationalSecurityAgency/ghidra) is an open-source software reverse engineering framework created and maintained by the National Security Agency Research Directorate. capa + Ghidra brings capa’s detection capabilities directly to Ghidra’s user interface helping speed up your reverse engineering tasks by identifying what parts of a program suggest interesting behavior, such as setting a registry value. You can execute the included Python 3 scripts [capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_explorer.py) or [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) to run capa’s analysis and view the results in Ghidra. You may be asking yourself, “Python 3 scripts in Ghidra?”. You read that correctly. This integration is written entirely in Python 3 and relies on [Ghidrathon]( https://github.com/mandiant/ghidrathon), an open source Ghidra extension that adds Python 3 scripting to Ghidra.
+
+Check out our capa + Ghidra blog posts:
+* [Riding Dragons: capa Harnesses Ghidra](https://www.mandiant.com/resources/blog/capa-harnesses-ghidra)
+
+## UI Integration
+[capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_explorer.py) renders capa results in Ghidra's UI to help you quickly navigate them. This includes adding matched functions to Ghidra’s Symbol Tree and Bookmarks windows and adding comments to functions that indicate matched capabilities and features. You can execute this script using Ghidra’s Script Manager window.
+
+### Symbol Tree Window
+Matched functions are added to Ghidra's Symbol Tree window under a custom namespace that maps to the capabilities' [capa namespace](https://github.com/mandiant/capa-rules/blob/master/doc/format.md#rule-namespace).
+<div align="center">
+    <img src="https://github.com/mandiant/capa/assets/66766340/eeae33f4-99d4-42dc-a5e8-4c1b8c661492" width=300>
+</div>
+
+### Comments
+
+Comments are added at the beginning of matched functions indicating matched capabilities and inline comments are added to functions indicating matched features. You can view these comments in Ghidra’s Disassembly Listing and Decompile windows.
+<div align="center">
+    <img src="https://github.com/mandiant/capa/assets/66766340/bb2b4170-7fd4-45fc-8c7b-ff8f2e2f101b" width=1000>
+</div>
+
+### Bookmarks
+
+Bookmarks are added to functions that matched a capability that is mapped to a MITRE ATT&CK and/or Malware Behavior Catalog (MBC) technique. You can view these bookmarks in Ghidra's Bookmarks window.
+<div align="center">
+    <img src="https://github.com/mandiant/capa/assets/66766340/7f9a66a9-7be7-4223-91c6-4b8fc4651336" width=825>
+</div>
+
+## Text-based Integration
+
+[capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) outputs text-based capa results that mirror the output of capa’s standalone tool. You can execute this script using Ghidra’s Script Manager and view its output in Ghidra’s Console window.
+
+<div align="center">
+  <img src="../../doc/img/ghidra_script_mngr_output.png" width=700>
+</div>
+
+You can also execute [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) using Ghidra's Headless Analyzer to view its output in a terminal window.
+
+<div align="center">
+  <img src="../../doc/img/ghidra_headless_analyzer.png">
+</div>
+
+# Getting Started
+
+## Requirements
+
+| Tool | Version | Source |
+|------------|---------|--------|
+| capa | `>= 7.0.0` | https://github.com/mandiant/capa/releases |
+| Ghidrathon | `>= 3.0.0` | https://github.com/mandiant/Ghidrathon/releases |
+| Ghidra | `>= 10.3.2` | https://github.com/NationalSecurityAgency/ghidra/releases |
+| Python | `>= 3.10.0` | https://www.python.org/downloads |
+
+## Installation
+
+**Note**: capa + Ghidra relies on [Ghidrathon]( https://github.com/mandiant/ghidrathon) to execute Python 3 code in Ghidra. You must first install and configure Ghidrathon using the [steps outlined in its README]( https://github.com/mandiant/ghidrathon?tab=readme-ov-file#installing-ghidrathon). Then, you must use the Python 3 interpreter that you configured with Ghidrathon to complete the following steps:
+
+1. Install capa and its dependencies from PyPI using the following command:
 ```bash
-$ capa -b ghidra Practical\ Malware\ Analysis\ Lab\ 01-01.exe_
-┌──────────┬──────────────────────────────────────────────────────────────────────────────────────────────────────┐
-│ md5      │ bb7425b82141a1c0f7d60e5106676bb1                                                                     │
-│ sha1     │                                                                                                      │
-│ sha256   │ 58898bd42c5bd3bf9b1389f0eee5b39cd59180e8370eb9ea838a0b327bd6fe47                                     │
-│ analysis │ static                                                                                               │
-│ os       │ windows                                                                                              │
-│ format   │ pe                                                                                                   │
-│ arch     │ i386                                                                                                 │
-│ path     │ ~/Documents/capa/tests/data/Practical Malware Analysis Lab 01-01.exe_                                │
-└──────────┴──────────────────────────────────────────────────────────────────────────────────────────────────────┘
-┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
-┃ ATT&CK Tactic                      ┃ ATT&CK Technique                                            ┃
-┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
-│ DISCOVERY                          │ File and Directory Discovery [T1083]                        │
-└────────────────────────────────────┴─────────────────────────────────────────────────────────────┘
-┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
-┃ MBC Objective                      ┃ MBC Behavior                                                ┃
-┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
-│ DISCOVERY                          │ File and Directory Discovery [E1083]                        │
-│ FILE SYSTEM                        │ Copy File [C0045]                                           │
-│                                    │ Read File [C0051]                                           │
-│ PROCESS                            │ Terminate Process [C0018]                                   │
-└────────────────────────────────────┴─────────────────────────────────────────────────────────────┘
-┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
-┃ Capability                                     ┃ Namespace                                       ┃
-┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
-│ copy file                                      │ host-interaction/file-system/copy               │
-│ enumerate files recursively                    │ host-interaction/file-system/files/list         │
-│ read file via mapping (2 matches)              │ host-interaction/file-system/read               │
-│ terminate process (2 matches)                  │ host-interaction/process/terminate              │
-│ resolve function by parsing PE exports         │ load-code/pe                                    │
-└────────────────────────────────────────────────┴─────────────────────────────────────────────────┘
+$ pip install flare-capa
 ```
 
-## getting started
-
-### requirements
-
-- [Ghidra](https://github.com/NationalSecurityAgency/ghidra) >= 12.0 must be installed and available via the `GHIDRA_INSTALL_DIR` environment variable.
-
-#### standalone binary (recommended)
-
-The capa [standalone binary](https://github.com/mandiant/capa/releases) is the preferred way to run capa with the Ghidra backend.
-Although the binary does not bundle the Java environment or Ghidra itself, it will dynamically load them at runtime.
-
-#### python package
-
-You can also use the Ghidra backend with the capa Python package by installing `flare-capa` with the `ghidra` extra.
-
+2. Download and extract the [official capa rules](https://github.com/mandiant/capa-rules/releases) that match the capa version you have installed. You can use the following command to view the version of capa you have installed:
 ```bash
-$ pip install "flare-capa[ghidra]"
+$ pip show flare-capa
+OR
+$ capa --version
 ```
 
-### usage
+3. Copy [capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_explorer.py) and [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) to your `ghidra_scripts` directory or manually add the parent directory of each script using Ghidra’s Script Manager.
 
-To use the Ghidra backend, specify it with the `-b` or `--backend` flag:
+## Usage
 
+You can execute [capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_explorer.py) and [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) using Ghidra’s Script Manager. [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) can also be executed using Ghidra's Headless Analyzer.
+
+### Execution using Ghidra’s Script Manager
+
+You can execute [capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_explorer.py) and [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) using Ghidra's Script Manager as follows:
+1. Navigate to `Window > Script Manager`
+2. Expand the `Python 3 > capa` category
+3. Double-click a script to execute it
+
+Both scripts ask you to provide the path of your capa rules directory (see installation step 2). [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) also has you choose one of `default`, `verbose`, and `vverbose` output formats which mirror the output formats of capa’s standalone tool.
+
+### Execution using Ghidra’s Headless Analyzer
+
+You can execute [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) using Ghidra’s Headless Analyzer by invoking the `analyzeHeadless` script included with Ghidra in its `support` directory. The following arguments must be provided:
+
+| Argument | Description |
+|----|----|
+|`<project_path>`| Path to Ghidra project|
+| `<project_name>`| Name of Ghidra Project|
+| `-Process <sample_name>` OR `-Import <sample_path>`| Name of sample `<sample_name>` already imported into `<project_name>` OR absolute path of sample `<sample_path>` to import into `<project_name>`|
+| `-ScriptPath <script_path>`| OPTIONAL parent directory `<script_path>` of [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py)|
+| `-PostScript capa_ghidra.py`| Execute [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) after Ghidra analysis|
+| `"<script_args>"`| Quoted string `"<script_args>"` containing script arguments passed to [capa_ghidra.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_ghidra.py) that must specify a capa rules path and optionally the output format (`--verbose`, `--vverbose`, `--json`) – you can specify `”help”` to view the script’s help message |
+
+The following is an example of combining these arguments into a single `analyzeHeadless` script command:
 ```bash
-$ capa -b ghidra /path/to/sample
+$ analyzeHeadless /home/wumbo/demo demo -Import /home/wumbo/capa/tests/data/Practical\ Malware\ Analysis\ Lab\ 01-01.dll_ -PostScript capa_ghidra.py "/home/wumbo/capa/rules --verbose"
 ```
-
-capa will:
-1.  Initialize a headless Ghidra instance.
-2.  Create a temporary project.
-3.  Import and analyze the sample.
-4.  Extract features and match rules.
-5.  Clean up the temporary project.
-
-**Note:** The first time you run this, it may take a few moments to initialize the Ghidra environment.

capa/ghidra/capa_explorer.py

@@ -1,3 +1,7 @@
+# Run capa against loaded Ghidra database and render results in Ghidra UI
+# @author Colton Gabertan (gabertan.colton@gmail.com)
+# @category Python 3.capa
+
 # Copyright 2024 Google LLC
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -12,63 +16,36 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# Run capa against loaded Ghidra database and render results in Ghidra UI
-
-# @author Colton Gabertan (gabertan.colton@gmail.com)
-# @category capa
-# @runtime PyGhidra
-
+import sys
 import json
 import logging
 import pathlib
 from typing import Any
 
-from java.util import ArrayList
-from ghidra.util import Msg
 from ghidra.app.cmd.label import AddLabelCmd, CreateNamespacesCmd
-from ghidra.util.exception import CancelledException
-from ghidra.program.flatapi import FlatProgramAPI
 from ghidra.program.model.symbol import Namespace, SourceType, SymbolType
 
 import capa
 import capa.main
 import capa.rules
-import capa.version
 import capa.render.json
 import capa.ghidra.helpers
 import capa.capabilities.common
-import capa.features.extractors.ghidra.context
 import capa.features.extractors.ghidra.extractor
 
 logger = logging.getLogger("capa_explorer")
 
 
-def show_monitor_message(msg):
-    capa.ghidra.helpers.get_monitor().checkCanceled()
-    capa.ghidra.helpers.get_monitor().setMessage(msg)
-
-
-def show_error(msg):
-    Msg.showError(None, None, "capa explorer", msg)
-
-
-def show_warn(msg):
-    Msg.showWarn(None, None, "capa explorer", msg)
-
-
-def show_info(msg):
-    Msg.showInfo(None, None, "capa explorer", msg)
-
-
 def add_bookmark(addr, txt, category="CapaExplorer"):
     """create bookmark at addr"""
-    capa.ghidra.helpers.get_current_program().getBookmarkManager().setBookmark(addr, "Info", category, txt)
+    currentProgram().getBookmarkManager().setBookmark(addr, "Info", category, txt)  # type: ignore [name-defined] # noqa: F821
 
 
 def create_namespace(namespace_str):
     """create new Ghidra namespace for each capa namespace"""
+
     cmd = CreateNamespacesCmd(namespace_str, SourceType.USER_DEFINED)
-    cmd.applyTo(capa.ghidra.helpers.get_current_program())
+    cmd.applyTo(currentProgram())  # type: ignore [name-defined] # noqa: F821
     return cmd.getNamespace()
 
 
@@ -76,7 +53,7 @@ def create_label(ghidra_addr, name, capa_namespace):
     """custom label cmd to overlay symbols under capa-generated namespaces"""
 
     # prevent duplicate labels under the same capa-generated namespace
-    symbol_table = capa.ghidra.helpers.get_current_program().getSymbolTable()
+    symbol_table = currentProgram().getSymbolTable()  # type: ignore [name-defined] # noqa: F821
     for sym in symbol_table.getSymbols(ghidra_addr):
         if sym.getName(True) == capa_namespace.getName(True) + Namespace.DELIMITER + name:
             return
@@ -84,7 +61,7 @@ def create_label(ghidra_addr, name, capa_namespace):
     # create SymbolType.LABEL at addr
     # prioritize capa-generated namespace (duplicate match @ new addr), else put under global Ghidra one (new match)
     cmd = AddLabelCmd(ghidra_addr, name, True, SourceType.USER_DEFINED)
-    cmd.applyTo(capa.ghidra.helpers.get_current_program())
+    cmd.applyTo(currentProgram())  # type: ignore [name-defined] # noqa: F821
 
     # assign new match overlay label to capa-generated namespace
     cmd.getSymbol().setNamespace(capa_namespace)
@@ -115,8 +92,8 @@ class CapaMatchData:
             return
 
         for key in self.matches.keys():
-            addr = capa.ghidra.helpers.get_flat_api().toAddr(hex(key))
-            func = capa.ghidra.helpers.get_flat_api().getFunctionContaining(addr)
+            addr = toAddr(hex(key))  # type: ignore [name-defined] # noqa: F821
+            func = getFunctionContaining(addr)  # type: ignore [name-defined] # noqa: F821
 
             # bookmark & tag MITRE ATT&CK tactics & MBC @ function scope
             if func is not None:
@@ -140,160 +117,140 @@ class CapaMatchData:
 
     def set_plate_comment(self, ghidra_addr):
         """set plate comments at matched functions"""
-        comment = capa.ghidra.helpers.get_flat_api().getPlateComment(ghidra_addr)
+        comment = getPlateComment(ghidra_addr)  # type: ignore [name-defined] # noqa: F821
         rule_path = self.namespace.replace(Namespace.DELIMITER, "/")
         # 2 calls to avoid duplicate comments via subsequent script runs
         if comment is None:
             # first comment @ function
             comment = rule_path + "\n"
-            capa.ghidra.helpers.get_flat_api().setPlateComment(ghidra_addr, comment)
+            setPlateComment(ghidra_addr, comment)  # type: ignore [name-defined] # noqa: F821
         elif rule_path not in comment:
             comment = comment + rule_path + "\n"
-            capa.ghidra.helpers.get_flat_api().setPlateComment(ghidra_addr, comment)
+            setPlateComment(ghidra_addr, comment)  # type: ignore [name-defined] # noqa: F821
         else:
             return
 
     def set_pre_comment(self, ghidra_addr, sub_type, description):
         """set pre comments at subscoped matches of main rules"""
-        comment = capa.ghidra.helpers.get_flat_api().getPreComment(ghidra_addr)
+        comment = getPreComment(ghidra_addr)  # type: ignore [name-defined] # noqa: F821
         if comment is None:
             comment = "capa: " + sub_type + "(" + description + ")" + ' matched in "' + self.capability + '"\n'
-            capa.ghidra.helpers.get_flat_api().setPreComment(ghidra_addr, comment)
+            setPreComment(ghidra_addr, comment)  # type: ignore [name-defined] # noqa: F821
         elif self.capability not in comment:
             comment = (
                 comment + "capa: " + sub_type + "(" + description + ")" + ' matched in "' + self.capability + '"\n'
             )
-            capa.ghidra.helpers.get_flat_api().setPreComment(ghidra_addr, comment)
+            setPreComment(ghidra_addr, comment)  # type: ignore [name-defined] # noqa: F821
         else:
             return
 
-    def label_matches(self, do_namespaces, do_comments):
+    def label_matches(self):
         """label findings at function scopes and comment on subscope matches"""
-        capa_namespace = None
-        if do_namespaces:
-            capa_namespace = create_namespace(self.namespace)
-
-        symbol_table = capa.ghidra.helpers.get_current_program().getSymbolTable()
+        capa_namespace = create_namespace(self.namespace)
+        symbol_table = currentProgram().getSymbolTable()  # type: ignore [name-defined] # noqa: F821
 
         # handle function main scope of matched rule
         # these will typically contain further matches within
         if self.scope == "function":
             for addr in self.matches.keys():
-                ghidra_addr = capa.ghidra.helpers.get_flat_api().toAddr(hex(addr))
+                ghidra_addr = toAddr(hex(addr))  # type: ignore [name-defined] # noqa: F821
 
                 # classify new function label under capa-generated namespace
-                if do_namespaces:
-                    sym = symbol_table.getPrimarySymbol(ghidra_addr)
-                    if sym is not None:
-                        if sym.getSymbolType() == SymbolType.FUNCTION:
-                            create_label(ghidra_addr, sym.getName(), capa_namespace)
+                sym = symbol_table.getPrimarySymbol(ghidra_addr)
+                if sym is not None:
+                    if sym.getSymbolType() == SymbolType.FUNCTION:
+                        create_label(ghidra_addr, sym.getName(), capa_namespace)
+                        self.set_plate_comment(ghidra_addr)
 
-                if do_comments:
-                    self.set_plate_comment(ghidra_addr)
+                    # parse the corresponding nodes, and pre-comment subscope matched features
+                    # under the encompassing function(s)
+                    for sub_match in self.matches.get(addr):
+                        for loc, node in sub_match.items():
+                            sub_ghidra_addr = toAddr(hex(loc))  # type: ignore [name-defined] # noqa: F821
+                            if sub_ghidra_addr == ghidra_addr:
+                                # skip duplicates
+                                continue
 
-                # parse the corresponding nodes, and pre-comment subscope matched features
-                # under the encompassing function(s)
-                for sub_match in self.matches.get(addr):
-                    for loc, node in sub_match.items():
-                        sub_ghidra_addr = capa.ghidra.helpers.get_flat_api().toAddr(hex(loc))
-                        if sub_ghidra_addr == ghidra_addr:
-                            # skip duplicates
-                            continue
-
-                        # precomment subscope matches under the function
-                        if node != {} and do_comments:
-                            for sub_type, description in parse_node(node):
-                                self.set_pre_comment(sub_ghidra_addr, sub_type, description)
+                            # precomment subscope matches under the function
+                            if node != {}:
+                                for sub_type, description in parse_node(node):
+                                    self.set_pre_comment(sub_ghidra_addr, sub_type, description)
         else:
             # resolve the encompassing function for the capa namespace
             # of non-function scoped main matches
             for addr in self.matches.keys():
-                ghidra_addr = capa.ghidra.helpers.get_flat_api().toAddr(hex(addr))
+                ghidra_addr = toAddr(hex(addr))  # type: ignore [name-defined] # noqa: F821
 
                 # basic block / insn scoped main matches
                 # Ex. See "Create Process on Windows" Rule
-                func = capa.ghidra.helpers.get_flat_api().getFunctionContaining(ghidra_addr)
+                func = getFunctionContaining(ghidra_addr)  # type: ignore [name-defined] # noqa: F821
                 if func is not None:
                     func_addr = func.getEntryPoint()
-                    if do_namespaces:
-                        create_label(func_addr, func.getName(), capa_namespace)
-                    if do_comments:
-                        self.set_plate_comment(func_addr)
+                    create_label(func_addr, func.getName(), capa_namespace)
+                    self.set_plate_comment(func_addr)
 
                 # create subscope match precomments
                 for sub_match in self.matches.get(addr):
                     for loc, node in sub_match.items():
-                        sub_ghidra_addr = capa.ghidra.helpers.get_flat_api().toAddr(hex(loc))
+                        sub_ghidra_addr = toAddr(hex(loc))  # type: ignore [name-defined] # noqa: F821
 
                         if node != {}:
                             if func is not None:
                                 # basic block/ insn scope under resolved function
-                                if do_comments:
-                                    for sub_type, description in parse_node(node):
-                                        self.set_pre_comment(sub_ghidra_addr, sub_type, description)
+                                for sub_type, description in parse_node(node):
+                                    self.set_pre_comment(sub_ghidra_addr, sub_type, description)
                             else:
                                 # this would be a global/file scoped main match
                                 # try to resolve the encompassing function via the subscope match, instead
                                 # Ex. "run as service" rule
-                                sub_func = capa.ghidra.helpers.get_flat_api().getFunctionContaining(sub_ghidra_addr)
+                                sub_func = getFunctionContaining(sub_ghidra_addr)  # type: ignore [name-defined] # noqa: F821
                                 if sub_func is not None:
                                     sub_func_addr = sub_func.getEntryPoint()
                                     # place function in capa namespace & create the subscope match label in Ghidra's global namespace
-                                    if do_namespaces:
-                                        create_label(sub_func_addr, sub_func.getName(), capa_namespace)
-                                    if do_comments:
-                                        self.set_plate_comment(sub_func_addr)
-
-                                    if do_comments:
-                                        for sub_type, description in parse_node(node):
-                                            self.set_pre_comment(sub_ghidra_addr, sub_type, description)
+                                    create_label(sub_func_addr, sub_func.getName(), capa_namespace)
+                                    self.set_plate_comment(sub_func_addr)
+                                    for sub_type, description in parse_node(node):
+                                        self.set_pre_comment(sub_ghidra_addr, sub_type, description)
                                 else:
                                     # addr is in some other file section like .data
                                     # represent this location with a label symbol under the capa namespace
                                     # Ex. See "Reference Base64 String" rule
-                                    if do_namespaces:
-                                        for _sub_type, _description in parse_node(node):
-                                            # in many cases, these will be ghidra-labeled data, so just add the existing
-                                            # label symbol to the capa namespace
-                                            for sym in symbol_table.getSymbols(sub_ghidra_addr):
-                                                if sym.getSymbolType() == SymbolType.LABEL:
-                                                    sym.setNamespace(capa_namespace)
-                                    if do_comments:
-                                        for sub_type, description in parse_node(node):
-                                            self.set_pre_comment(sub_ghidra_addr, sub_type, description)
+                                    for sub_type, description in parse_node(node):
+                                        # in many cases, these will be ghidra-labeled data, so just add the existing
+                                        # label symbol to the capa namespace
+                                        for sym in symbol_table.getSymbols(sub_ghidra_addr):
+                                            if sym.getSymbolType() == SymbolType.LABEL:
+                                                sym.setNamespace(capa_namespace)
+                                        self.set_pre_comment(sub_ghidra_addr, sub_type, description)
 
 
 def get_capabilities():
-    rules_dir = ""
-
-    show_monitor_message(f"requesting capa {capa.version.__version__} rules directory")
-    selected_dir = askDirectory(f"choose capa {capa.version.__version__} rules directory", "Ok")  # type: ignore [name-defined] # noqa: F821
-
-    if selected_dir:
-        rules_dir = selected_dir.getPath()
+    rules_dir: str = ""
+    try:
+        selected_dir = askDirectory("Choose capa rules directory", "Ok")  # type: ignore [name-defined] # noqa: F821
+        if selected_dir:
+            rules_dir = selected_dir.getPath()
+    except RuntimeError:
+        # RuntimeError thrown when user selects "Cancel"
+        pass
 
     if not rules_dir:
-        raise CancelledException
+        logger.info("You must choose a capa rules directory before running capa.")
+        return ""  # return empty str to avoid handling both int and str types
 
     rules_path: pathlib.Path = pathlib.Path(rules_dir)
+    logger.info("running capa using rules from %s", str(rules_path))
 
-    show_monitor_message(f"loading rules from {rules_path}")
     rules = capa.rules.get_rules([rules_path])
-
-    show_monitor_message("collecting binary metadata")
     meta = capa.ghidra.helpers.collect_metadata([rules_path])
-
-    show_monitor_message("running capa analysis")
     extractor = capa.features.extractors.ghidra.extractor.GhidraFeatureExtractor()
+
     capabilities = capa.capabilities.common.find_capabilities(rules, extractor, True)
 
-    show_monitor_message("checking for static limitations")
     if capa.capabilities.common.has_static_limitation(rules, capabilities, is_standalone=False):
-        show_warn(
-            "capa explorer encountered warnings during analysis. Please check the console output for more information.",
-        )
+        popup("capa explorer encountered warnings during analysis. Please check the console output for more information.")  # type: ignore [name-defined] # noqa: F821
+        logger.info("capa encountered warnings during analysis")
 
-    show_monitor_message("rendering results")
     return capa.render.json.render(meta, rules, capabilities.matches)
 
 
@@ -371,12 +328,12 @@ def parse_json(capa_data):
             # this requires the correct delimiter used by Ghidra
             # Ex. 'communication/named-pipe/create/create pipe' -> capa::communication::named-pipe::create::create-pipe
             namespace_str = Namespace.DELIMITER.join(meta["namespace"].split("/"))
-            namespace = "capa_explorer" + Namespace.DELIMITER + namespace_str + fmt_rule
+            namespace = "capa" + Namespace.DELIMITER + namespace_str + fmt_rule
         else:
             # lib rules via the official rules repo will not contain data
             # for the "namespaces" key, so format using rule itself
             # Ex. 'contain loop' -> capa::lib::contain-loop
-            namespace = "capa_explorer" + Namespace.DELIMITER + "lib" + fmt_rule
+            namespace = "capa" + Namespace.DELIMITER + "lib" + fmt_rule
 
         yield CapaMatchData(namespace, scope, rule, rule_matches, attack, mbc)
 
@@ -385,79 +342,44 @@ def main():
     logging.basicConfig(level=logging.INFO)
     logging.getLogger().setLevel(logging.INFO)
 
-    choices = ["namespaces", "bookmarks", "comments"]
-    # use ArrayList to resolve ambiguous askChoices overloads (List vs List, List) in PyGhidra
-    choices_java = ArrayList()
-    for c in choices:
-        choices_java.add(c)
+    if isRunningHeadless():  # type: ignore [name-defined] # noqa: F821
+        logger.error("unsupported Ghidra execution mode")
+        return capa.main.E_UNSUPPORTED_GHIDRA_EXECUTION_MODE
 
-    choice_labels = [
-        'add "capa_explorer" namespace for matched functions',
-        "add bookmarks for matched functions",
-        "add comments to matched functions",
-    ]
-    # use ArrayList to resolve ambiguous askChoices overloads (List vs List, List) in PyGhidra
-    choice_labels_java = ArrayList()
-    for c in choice_labels:
-        choice_labels_java.add(c)
-
-    selected = list(askChoices("capa explorer", "select actions:", choices_java, choice_labels_java))  # type: ignore [name-defined] # noqa: F821
-
-    do_namespaces = "namespaces" in selected
-    do_comments = "comments" in selected
-    do_bookmarks = "bookmarks" in selected
-
-    if not any((do_namespaces, do_comments, do_bookmarks)):
-        raise CancelledException("no actions selected")
-
-    # initialize the context for the extractor/helpers
-    capa.features.extractors.ghidra.context.set_context(
-        currentProgram,  # type: ignore [name-defined] # noqa: F821
-        FlatProgramAPI(currentProgram),  # type: ignore [name-defined] # noqa: F821
-        monitor,  # type: ignore [name-defined] # noqa: F821
-    )
-
-    show_monitor_message("checking supported Ghidra version")
     if not capa.ghidra.helpers.is_supported_ghidra_version():
-        show_error("unsupported Ghidra version")
+        logger.error("unsupported Ghidra version")
         return capa.main.E_UNSUPPORTED_GHIDRA_VERSION
 
-    show_monitor_message("checking supported file type")
     if not capa.ghidra.helpers.is_supported_file_type():
-        show_error("unsupported file type")
+        logger.error("unsupported file type")
         return capa.main.E_INVALID_FILE_TYPE
 
-    show_monitor_message("checking supported file architecture")
     if not capa.ghidra.helpers.is_supported_arch_type():
-        show_error("unsupported file architecture")
+        logger.error("unsupported file architecture")
         return capa.main.E_INVALID_FILE_ARCH
 
     # capa_data will always contain {'meta':..., 'rules':...}
     # if the 'rules' key contains no values, then there were no matches
     capa_data = json.loads(get_capabilities())
     if capa_data.get("rules") is None:
-        show_info("capa explorer found no matches.")
+        logger.info("capa explorer found no matches")
+        popup("capa explorer found no matches.")  # type: ignore [name-defined] # noqa: F821
         return capa.main.E_EMPTY_REPORT
 
-    show_monitor_message("processing matches")
     for item in parse_json(capa_data):
-        if do_bookmarks:
-            show_monitor_message("adding bookmarks")
-            item.bookmark_functions()
-        if do_namespaces or do_comments:
-            show_monitor_message("adding labels")
-            item.label_matches(do_namespaces, do_comments)
-
-    show_info("capa explorer analysis complete.")
-
+        item.bookmark_functions()
+        item.label_matches()
+    logger.info("capa explorer analysis complete")
+    popup("capa explorer analysis complete.\nPlease see results in the Bookmarks Window and Namespaces section of the Symbol Tree Window.")  # type: ignore [name-defined] # noqa: F821
     return 0
 
 
 if __name__ == "__main__":
-    try:
-        if main() != 0:
-            show_error(
-                "capa explorer encountered errors during analysis. Please check the console output for more information.",
-            )
-    except CancelledException:
-        show_info("capa explorer analysis cancelled.")
+    if sys.version_info < (3, 10):
+        from capa.exceptions import UnsupportedRuntimeError
+
+        raise UnsupportedRuntimeError("This version of capa can only be used with Python 3.10+")
+    exit_code = main()
+    if exit_code != 0:
+        popup("capa explorer encountered errors during analysis. Please check the console output for more information.")  # type: ignore [name-defined] # noqa: F821
+    sys.exit(exit_code)

capa/ghidra/capa_ghidra.py

@@ -0,0 +1,174 @@
+# Run capa against loaded Ghidra database and render results in Ghidra Console window
+# @author Mike Hunhoff (mehunhoff@google.com)
+# @category Python 3.capa
+
+# Copyright 2023 Google LLC
+#
+# 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
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# 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 logging
+import pathlib
+import argparse
+
+import capa
+import capa.main
+import capa.rules
+import capa.ghidra.helpers
+import capa.render.default
+import capa.capabilities.common
+import capa.features.extractors.ghidra.extractor
+
+logger = logging.getLogger("capa_ghidra")
+
+
+def run_headless():
+    parser = argparse.ArgumentParser(description="The FLARE team's open-source tool to integrate capa with Ghidra.")
+
+    parser.add_argument(
+        "rules",
+        type=str,
+        help="path to rule file or directory",
+    )
+    parser.add_argument(
+        "-v", "--verbose", action="store_true", help="enable verbose result document (no effect with --json)"
+    )
+    parser.add_argument(
+        "-vv", "--vverbose", action="store_true", help="enable very verbose result document (no effect with --json)"
+    )
+    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")
+    parser.add_argument("-j", "--json", action="store_true", help="emit JSON instead of text")
+
+    script_args = list(getScriptArgs())  # type: ignore [name-defined] # noqa: F821
+    if not script_args or len(script_args) > 1:
+        script_args = []
+    else:
+        script_args = script_args[0].split()
+        for idx, arg in enumerate(script_args):
+            if arg.lower() == "help":
+                script_args[idx] = "--help"
+
+    args = parser.parse_args(args=script_args)
+
+    if args.quiet:
+        logging.basicConfig(level=logging.WARNING)
+        logging.getLogger().setLevel(logging.WARNING)
+    elif args.debug:
+        logging.basicConfig(level=logging.DEBUG)
+        logging.getLogger().setLevel(logging.DEBUG)
+    else:
+        logging.basicConfig(level=logging.INFO)
+        logging.getLogger().setLevel(logging.INFO)
+
+    logger.debug("running in Ghidra headless mode")
+
+    rules_path = pathlib.Path(args.rules)
+
+    logger.debug("rule path: %s", rules_path)
+    rules = capa.rules.get_rules([rules_path])
+
+    meta = capa.ghidra.helpers.collect_metadata([rules_path])
+    extractor = capa.features.extractors.ghidra.extractor.GhidraFeatureExtractor()
+
+    capabilities = capa.capabilities.common.find_capabilities(rules, extractor, False)
+
+    meta.analysis.feature_counts = capabilities.feature_counts
+    meta.analysis.library_functions = capabilities.library_functions
+    meta.analysis.layout = capa.loader.compute_layout(rules, extractor, capabilities.matches)
+
+    if capa.capabilities.common.has_static_limitation(rules, capabilities, is_standalone=True):
+        logger.info("capa encountered warnings during analysis")
+
+    if args.json:
+        print(capa.render.json.render(meta, rules, capabilities.matches))  # noqa: T201
+    elif args.vverbose:
+        print(capa.render.vverbose.render(meta, rules, capabilities.matches))  # noqa: T201
+    elif args.verbose:
+        print(capa.render.verbose.render(meta, rules, capabilities.matches))  # noqa: T201
+    else:
+        print(capa.render.default.render(meta, rules, capabilities.matches))  # noqa: T201
+
+    return 0
+
+
+def run_ui():
+    logging.basicConfig(level=logging.INFO)
+    logging.getLogger().setLevel(logging.INFO)
+
+    rules_dir: str = ""
+    try:
+        selected_dir = askDirectory("Choose capa rules directory", "Ok")  # type: ignore [name-defined] # noqa: F821
+        if selected_dir:
+            rules_dir = selected_dir.getPath()
+    except RuntimeError:
+        # RuntimeError thrown when user selects "Cancel"
+        pass
+
+    if not rules_dir:
+        logger.info("You must choose a capa rules directory before running capa.")
+        return capa.main.E_MISSING_RULES
+
+    verbose = askChoice(  # type: ignore [name-defined] # noqa: F821
+        "capa output verbosity", "Choose capa output verbosity", ["default", "verbose", "vverbose"], "default"
+    )
+
+    rules_path: pathlib.Path = pathlib.Path(rules_dir)
+    logger.info("running capa using rules from %s", str(rules_path))
+
+    rules = capa.rules.get_rules([rules_path])
+
+    meta = capa.ghidra.helpers.collect_metadata([rules_path])
+    extractor = capa.features.extractors.ghidra.extractor.GhidraFeatureExtractor()
+
+    capabilities = capa.capabilities.common.find_capabilities(rules, extractor, True)
+
+    meta.analysis.feature_counts = capabilities.feature_counts
+    meta.analysis.library_functions = capabilities.library_functions
+    meta.analysis.layout = capa.loader.compute_layout(rules, extractor, capabilities.matches)
+
+    if capa.capabilities.common.has_static_limitation(rules, capabilities, is_standalone=False):
+        logger.info("capa encountered warnings during analysis")
+
+    if verbose == "vverbose":
+        print(capa.render.vverbose.render(meta, rules, capabilities.matches))  # noqa: T201
+    elif verbose == "verbose":
+        print(capa.render.verbose.render(meta, rules, capabilities.matches))  # noqa: T201
+    else:
+        print(capa.render.default.render(meta, rules, capabilities.matches))  # noqa: T201
+
+    return 0
+
+
+def main():
+    if not capa.ghidra.helpers.is_supported_ghidra_version():
+        return capa.main.E_UNSUPPORTED_GHIDRA_VERSION
+
+    if not capa.ghidra.helpers.is_supported_file_type():
+        return capa.main.E_INVALID_FILE_TYPE
+
+    if not capa.ghidra.helpers.is_supported_arch_type():
+        return capa.main.E_INVALID_FILE_ARCH
+
+    if isRunningHeadless():  # type: ignore [name-defined] # noqa: F821
+        return run_headless()
+    else:
+        return run_ui()
+
+
+if __name__ == "__main__":
+    if sys.version_info < (3, 10):
+        from capa.exceptions import UnsupportedRuntimeError
+
+        raise UnsupportedRuntimeError("This version of capa can only be used with Python 3.10+")
+    sys.exit(main())

capa/ghidra/helpers.py

@@ -22,7 +22,6 @@ import capa.version
 import capa.features.common
 import capa.features.freeze
 import capa.render.result_document as rdoc
-import capa.features.extractors.ghidra.context as ghidra_context
 import capa.features.extractors.ghidra.helpers
 from capa.features.address import AbsoluteVirtualAddress
 
@@ -32,18 +31,6 @@ logger = logging.getLogger("capa")
 SUPPORTED_FILE_TYPES = ("Executable and Linking Format (ELF)", "Portable Executable (PE)", "Raw Binary")
 
 
-def get_current_program():
-    return ghidra_context.get_context().program
-
-
-def get_flat_api():
-    return ghidra_context.get_context().flat_api
-
-
-def get_monitor():
-    return ghidra_context.get_context().monitor
-
-
 class GHIDRAIO:
     """
     An object that acts as a file-like object,
@@ -61,12 +48,7 @@ class GHIDRAIO:
         self.offset = offset
 
     def read(self, size):
-        logger.debug(
-            "reading 0x%x bytes at 0x%x (ea: 0x%x)",
-            size,
-            self.offset,
-            get_current_program().getImageBase().add(self.offset).getOffset(),
-        )
+        logger.debug("reading 0x%x bytes at 0x%x (ea: 0x%x)", size, self.offset, currentProgram().getImageBase().add(self.offset).getOffset())  # type: ignore [name-defined] # noqa: F821
 
         if size > len(self.bytes_) - self.offset:
             logger.debug("cannot read 0x%x bytes at 0x%x (ea: BADADDR)", size, self.offset)
@@ -78,7 +60,7 @@ class GHIDRAIO:
         return
 
     def get_bytes(self):
-        file_bytes = get_current_program().getMemory().getAllFileBytes()[0]
+        file_bytes = currentProgram().getMemory().getAllFileBytes()[0]  # type: ignore [name-defined] # noqa: F821
 
         # getOriginalByte() allows for raw file parsing on the Ghidra side
         # other functions will fail as Ghidra will think that it's reading uninitialized memory
@@ -88,32 +70,21 @@ class GHIDRAIO:
 
 
 def is_supported_ghidra_version():
-    import ghidra.framework
-
-    version = ghidra.framework.Application.getApplicationVersion()
-    try:
-        # version format example: "11.1.2" or "11.4"
-        major, minor = map(int, version.split(".")[:2])
-        if major < 12:
-            logger.error("-" * 80)
-            logger.error(" Ghidra version %s is not supported.", version)
-            logger.error(" ")
-            logger.error(" capa requires Ghidra 12.0 or higher.")
-            logger.error("-" * 80)
-            return False
-    except ValueError:
-        logger.warning("could not parse Ghidra version: %s", version)
+    version = float(getGhidraVersion()[:4])  # type: ignore [name-defined] # noqa: F821
+    if version < 10.2:
+        warning_msg = "capa does not support this Ghidra version"
+        logger.warning(warning_msg)
+        logger.warning("Your Ghidra version is: %s. Supported versions are: Ghidra >= 10.2", version)
         return False
-
     return True
 
 
 def is_running_headless():
-    return True  # PyGhidra is always headless in this context
+    return isRunningHeadless()  # type: ignore [name-defined] # noqa: F821
 
 
 def is_supported_file_type():
-    file_info = get_current_program().getExecutableFormat()
+    file_info = currentProgram().getExecutableFormat()  # type: ignore [name-defined] # noqa: F821
     if file_info not in SUPPORTED_FILE_TYPES:
         logger.error("-" * 80)
         logger.error(" Input file does not appear to be a supported file type.")
@@ -128,7 +99,7 @@ def is_supported_file_type():
 
 
 def is_supported_arch_type():
-    lang_id = str(get_current_program().getLanguageID()).lower()
+    lang_id = str(currentProgram().getLanguageID()).lower()  # type: ignore [name-defined] # noqa: F821
 
     if not all((lang_id.startswith("x86"), any(arch in lang_id for arch in ("32", "64")))):
         logger.error("-" * 80)
@@ -141,18 +112,18 @@ def is_supported_arch_type():
 
 
 def get_file_md5():
-    return get_current_program().getExecutableMD5()
+    return currentProgram().getExecutableMD5()  # type: ignore [name-defined] # noqa: F821
 
 
 def get_file_sha256():
-    return get_current_program().getExecutableSHA256()
+    return currentProgram().getExecutableSHA256()  # type: ignore [name-defined] # noqa: F821
 
 
 def collect_metadata(rules: list[Path]):
     md5 = get_file_md5()
     sha256 = get_file_sha256()
 
-    info = get_current_program().getLanguageID().toString()
+    info = currentProgram().getLanguageID().toString()  # type: ignore [name-defined] # noqa: F821
     if "x86" in info and "64" in info:
         arch = "x86_64"
     elif "x86" in info and "32" in info:
@@ -160,11 +131,11 @@ def collect_metadata(rules: list[Path]):
     else:
         arch = "unknown arch"
 
-    format_name: str = get_current_program().getExecutableFormat()
+    format_name: str = currentProgram().getExecutableFormat()  # type: ignore [name-defined] # noqa: F821
     if "PE" in format_name:
         os = "windows"
     elif "ELF" in format_name:
-        with contextlib.closing(GHIDRAIO()) as f:
+        with contextlib.closing(capa.ghidra.helpers.GHIDRAIO()) as f:
             os = capa.features.extractors.elf.detect_elf_os(f)
     else:
         os = "unknown os"
@@ -177,18 +148,16 @@ def collect_metadata(rules: list[Path]):
             md5=md5,
             sha1="",
             sha256=sha256,
-            path=get_current_program().getExecutablePath(),
+            path=currentProgram().getExecutablePath(),  # type: ignore [name-defined] # noqa: F821
         ),
         flavor=rdoc.Flavor.STATIC,
         analysis=rdoc.StaticAnalysis(
-            format=get_current_program().getExecutableFormat(),
+            format=currentProgram().getExecutableFormat(),  # type: ignore [name-defined] # noqa: F821
             arch=arch,
             os=os,
             extractor="ghidra",
             rules=tuple(r.resolve().absolute().as_posix() for r in rules),
-            base_address=capa.features.freeze.Address.from_capa(
-                AbsoluteVirtualAddress(get_current_program().getImageBase().getOffset())
-            ),
+            base_address=capa.features.freeze.Address.from_capa(AbsoluteVirtualAddress(currentProgram().getImageBase().getOffset())),  # type: ignore [name-defined] # noqa: F821
             layout=rdoc.StaticLayout(
                 functions=(),
             ),

capa/ghidra/plugin/README.md

@@ -1,54 +0,0 @@
-<div align="center">
-    <img src="https://github.com/mandiant/capa/blob/master/doc/img/ghidra_backend_logo.png" width=240 height=125>
-</div>
-
-# capa explorer for Ghidra
-
-capa explorer for Ghidra brings capa’s detection capabilities directly to Ghidra’s user interface helping speed up your reverse engineering tasks by identifying what parts of a program suggest interesting behavior, such as setting a registry value. You can execute (via [PyGhidra](https://github.com/NationalSecurityAgency/ghidra/tree/master/Ghidra/Features/PyGhidra)) the script [capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/plugin/capa_explorer.py) using Ghidra’s Script Manager window to run capa’s analysis and view the results in Ghidra.
-
-## ui integration
-[capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/capa_explorer.py) renders capa results in Ghidra's UI to help you quickly navigate them. This includes adding matched functions to Ghidra’s Symbol Tree and Bookmarks windows and adding comments to functions that indicate matched capabilities and features. You can execute this script using Ghidra’s Script Manager window.
-
-### symbol tree window
-Matched functions are added to Ghidra's Symbol Tree window under a custom namespace that maps to the capabilities' [capa namespace](https://github.com/mandiant/capa-rules/blob/master/doc/format.md#rule-namespace).
-<div align="center">
-    <img src="https://github.com/mandiant/capa/assets/66766340/eeae33f4-99d4-42dc-a5e8-4c1b8c661492" width=300>
-</div>
-
-### comments
-
-Comments are added at the beginning of matched functions indicating matched capabilities and inline comments are added to functions indicating matched features. You can view these comments in Ghidra’s Disassembly Listing and Decompile windows.
-<div align="center">
-    <img src="https://github.com/mandiant/capa/assets/66766340/bb2b4170-7fd4-45fc-8c7b-ff8f2e2f101b" width=1000>
-</div>
-
-### bookmarks
-
-Bookmarks are added to functions that matched a capability that is mapped to a MITRE ATT&CK and/or Malware Behavior Catalog (MBC) technique. You can view these bookmarks in Ghidra's Bookmarks window.
-<div align="center">
-    <img src="https://github.com/mandiant/capa/assets/66766340/7f9a66a9-7be7-4223-91c6-4b8fc4651336" width=825>
-</div>
-
-# getting started
-
-## requirements
-
-- [Ghidra](https://github.com/NationalSecurityAgency/ghidra) >= 12.0 must be installed.
-- [flare-capa](https://pypi.org/project/flare-capa/) >= 10.0 must be installed (virtual environment recommended) with the `ghidra` extra (e.g., `pip install "flare-capa[ghidra]"`).
-- [capa rules](https://github.com/mandiant/capa-rules) must be downloaded for the version of capa you are using.
-
-## execution
-
-### 1. run Ghidra with PyGhidra
-You must start Ghidra using the `pyghidraRun` script provided in the support directory of your Ghidra installation to ensure the Python environment is correctly loaded. You should execute `pyghidraRun` from within the Python environment that you used to install capa.
-
-```bash
-<ghidra_install>/support/pyghidraRun
-```
-
-### 2. run capa_explorer.py
-1. Open your Ghidra project and CodeBrowser.
-2. Open the Script Manager.
-3. Add [capa_explorer.py](https://raw.githubusercontent.com/mandiant/capa/master/capa/ghidra/plugin/capa_explorer.py) to the script directories.
-4. Filter for capa and run the script.
-5. When prompted, select the directory containing the downloaded capa rules.

capa/helpers.py

@@ -96,7 +96,11 @@ def is_runtime_ida():
 
 
 def is_runtime_ghidra():
-    return importlib.util.find_spec("ghidra") is not None
+    try:
+        currentProgram  # type: ignore [name-defined] # noqa: F821
+    except NameError:
+        return False
+    return True
 
 
 def assert_never(value) -> NoReturn:

capa/loader.py

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import io
 import os
 import logging
 import datetime
@@ -23,24 +22,13 @@ from pathlib import Path
 from rich.console import Console
 from typing_extensions import assert_never
 
-import capa.perf
 import capa.rules
-import capa.engine
-import capa.helpers
 import capa.version
-import capa.render.json
-import capa.rules.cache
-import capa.render.default
-import capa.render.verbose
 import capa.features.common
 import capa.features.freeze as frz
-import capa.render.vverbose
 import capa.features.extractors
-import capa.render.result_document
 import capa.render.result_document as rdoc
 import capa.features.extractors.common
-import capa.features.extractors.base_extractor
-import capa.features.extractors.cape.extractor
 from capa.rules import RuleSet
 from capa.engine import MatchResults
 from capa.exceptions import UnsupportedOSError, UnsupportedArchError, UnsupportedFormatError
@@ -79,7 +67,6 @@ BACKEND_VMRAY = "vmray"
 BACKEND_FREEZE = "freeze"
 BACKEND_BINEXPORT2 = "binexport2"
 BACKEND_IDA = "ida"
-BACKEND_GHIDRA = "ghidra"
 
 
 class CorruptFile(ValueError):
@@ -339,12 +326,23 @@ def get_extractor(
         import capa.features.extractors.ida.extractor
 
         logger.debug("idalib: opening database...")
-        # idalib writes to stdout (ugh), so we have to capture that
-        # so as not to screw up structured output.
-        with capa.helpers.stdout_redirector(io.BytesIO()):
-            with console.status("analyzing program...", spinner="dots"):
-                if idapro.open_database(str(input_path), run_auto_analysis=True):
-                    raise RuntimeError("failed to analyze input file")
+        idapro.enable_console_messages(False)
+        with console.status("analyzing program...", spinner="dots"):
+            # we set the primary and secondary Lumina servers to 0.0.0.0 to disable Lumina,
+            # which sometimes provides bad names, including overwriting names from debug info.
+            #
+            # return values from open_database:
+            #   0 - Success (database not packed)
+            #   1 - Success (database was packed)
+            #   2 - User cancelled or 32-64 bit conversion failed
+            #   4 - Database initialization failed
+            #   -1 - Generic errors (database already open, auto-analysis failed, etc.)
+            #   -2 - User cancelled operation
+            ret = idapro.open_database(
+                str(input_path), run_auto_analysis=True, args="-Olumina:host=0.0.0.0 -Osecondary_lumina:host=0.0.0.0"
+            )
+            if ret not in (0, 1):
+                raise RuntimeError("failed to analyze input file")
 
             logger.debug("idalib: waiting for analysis...")
             ida_auto.auto_wait()
@@ -352,69 +350,6 @@ def get_extractor(
 
         return capa.features.extractors.ida.extractor.IdaFeatureExtractor()
 
-    elif backend == BACKEND_GHIDRA:
-        import pyghidra
-
-        with console.status("analyzing program...", spinner="dots"):
-            if not pyghidra.started():
-                pyghidra.start()
-
-            import capa.ghidra.helpers
-
-            if not capa.ghidra.helpers.is_supported_ghidra_version():
-                raise RuntimeError("unsupported Ghidra version")
-
-            import tempfile
-
-            tmpdir = tempfile.TemporaryDirectory()
-
-            project_cm = pyghidra.open_project(tmpdir.name, "CapaProject", create=True)
-            project = project_cm.__enter__()
-            try:
-                from ghidra.util.task import TaskMonitor
-
-                monitor = TaskMonitor.DUMMY
-
-                # Import file
-                loader = pyghidra.program_loader().project(project).source(str(input_path)).name(input_path.name)
-                with loader.load() as load_results:
-                    load_results.save(monitor)
-
-                # Open program
-                program, consumer = pyghidra.consume_program(project, "/" + input_path.name)
-
-                # Analyze
-                pyghidra.analyze(program, monitor)
-
-                from ghidra.program.flatapi import FlatProgramAPI
-
-                flat_api = FlatProgramAPI(program)
-
-                import capa.features.extractors.ghidra.context as ghidra_context
-
-                ghidra_context.set_context(program, flat_api, monitor)
-
-                # Wrapper to handle cleanup of program (consumer) and project
-                class GhidraContextWrapper:
-                    def __init__(self, project_cm, program, consumer):
-                        self.project_cm = project_cm
-                        self.program = program
-                        self.consumer = consumer
-
-                    def __exit__(self, exc_type, exc_val, exc_tb):
-                        self.program.release(self.consumer)
-                        self.project_cm.__exit__(exc_type, exc_val, exc_tb)
-
-                cm = GhidraContextWrapper(project_cm, program, consumer)
-
-            except Exception:
-                project_cm.__exit__(None, None, None)
-                tmpdir.cleanup()
-                raise
-
-        import capa.features.extractors.ghidra.extractor
-
-        return capa.features.extractors.ghidra.extractor.GhidraFeatureExtractor(ctx_manager=cm, tmpdir=tmpdir)
     else:
         raise ValueError("unexpected backend: " + backend)
 

capa/main.py

@@ -55,7 +55,6 @@ from capa.loader import (
     BACKEND_VMRAY,
     BACKEND_DOTNET,
     BACKEND_FREEZE,
-    BACKEND_GHIDRA,
     BACKEND_PEFILE,
     BACKEND_DRAKVUF,
     BACKEND_BINEXPORT2,
@@ -299,7 +298,6 @@ def install_common_args(parser, wanted=None):
             (BACKEND_BINJA, "Binary Ninja"),
             (BACKEND_DOTNET, ".NET"),
             (BACKEND_BINEXPORT2, "BinExport2"),
-            (BACKEND_GHIDRA, "Ghidra"),
             (BACKEND_FREEZE, "capa freeze"),
             (BACKEND_CAPE, "CAPE"),
             (BACKEND_DRAKVUF, "DRAKVUF"),
@@ -1107,26 +1105,14 @@ def ida_main():
 
 
 def ghidra_main():
-    from ghidra.program.flatapi import FlatProgramAPI
-
     import capa.rules
     import capa.ghidra.helpers
     import capa.render.default
-    import capa.features.extractors.ghidra.context
     import capa.features.extractors.ghidra.extractor
 
     logging.basicConfig(level=logging.INFO)
     logging.getLogger().setLevel(logging.INFO)
 
-    # These are provided by the Ghidra scripting environment
-    # but are not available when running standard python
-    # so we have to ignore the linting errors
-    program = currentProgram  # type: ignore [name-defined] # noqa: F821
-    monitor_ = monitor  # type: ignore [name-defined] # noqa: F821
-    flat_api = FlatProgramAPI(program)
-
-    capa.features.extractors.ghidra.context.set_context(program, flat_api, monitor_)
-
     logger.debug("-" * 80)
     logger.debug(" Using default embedded rules.")
     logger.debug(" ")

pyproject.toml

@@ -123,7 +123,7 @@ dev = [
     # These dependencies are not used in production environments
     # and should not conflict with other libraries/tooling.
     "pre-commit==4.5.0",
-    "pytest==9.0.2",
+    "pytest==8.0.0",
     "pytest-sugar==1.1.1",
     "pytest-instafail==0.5.0",
     "flake8==7.3.0",
@@ -139,8 +139,8 @@ dev = [
     "flake8-copyright==0.2.4",
     "ruff==0.14.7",
     "black==25.12.0",
-    "isort==7.0.0",
-    "mypy==1.19.1",
+    "isort==6.0.0",
+    "mypy==1.17.1",
     "mypy-protobuf==3.6.0",
     "PyGithub==2.6.0",
     "bump-my-version==1.2.4",
@@ -151,14 +151,14 @@ dev = [
     "types-psutil==7.1.3.20251202",
     "types_requests==2.32.0.20240712",
     "types-protobuf==6.32.1.20250918",
-    "deptry==0.24.0"
+    "deptry==0.23.0"
 ]
 build = [
     # Dev and build dependencies are not relaxed because
     # we want all developer environments to be consistent.
     # These dependencies are not used in production environments
     # and should not conflict with other libraries/tooling.
-    "pyinstaller==6.17.0",
+    "pyinstaller==6.16.0",
     "setuptools==80.9.0",
     "build==1.3.0"
 ]
@@ -171,9 +171,6 @@ scripts = [
     "sarif_om==1.0.4",
     "requests>=2.32.4",
 ]
-ghidra = [
-    "pyghidra>=3.0.0",
-]
 
 [tool.deptry]
 extend_exclude = [

requirements.txt

@@ -34,7 +34,6 @@ pydantic==2.12.4
 xmltodict==1.0.2
 pyelftools==0.32
 pygments==2.19.1
-pyghidra==3.0.0
 python-flirt==0.9.2
 pyyaml==6.0.2
 rich==14.2.0

tests/fixtures.py

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-
+import logging
 import contextlib
 import collections
 from pathlib import Path
@@ -20,7 +20,6 @@ from functools import lru_cache
 
 import pytest
 
-import capa.main
 import capa.features.file
 import capa.features.insn
 import capa.features.common
@@ -53,6 +52,7 @@ from capa.features.extractors.base_extractor import (
 )
 from capa.features.extractors.dnfile.extractor import DnfileFeatureExtractor
 
+logger = logging.getLogger(__name__)
 CD = Path(__file__).resolve().parent
 DOTNET_DIR = CD / "data" / "dotnet"
 DNFILE_TESTFILES = DOTNET_DIR / "dnfile-testfiles"
@@ -200,6 +200,77 @@ def get_binja_extractor(path: Path):
     return extractor
 
 
+# we can't easily cache this because the extractor relies on global state (the opened database)
+# which also has to be closed elsewhere. so, the idalib tests will just take a little bit to run.
+def get_idalib_extractor(path: Path):
+    import capa.features.extractors.ida.idalib as idalib
+
+    if not idalib.has_idalib():
+        raise RuntimeError("cannot find IDA idalib module.")
+
+    if not idalib.load_idalib():
+        raise RuntimeError("failed to load IDA idalib module.")
+
+    import idapro
+    import ida_auto
+
+    import capa.features.extractors.ida.extractor
+
+    logger.debug("idalib: opening database...")
+
+    idapro.enable_console_messages(False)
+
+    # load resource if explicitly needed for test
+    load_resource = ""
+    if "Lab 12-04.exe_" in path.name:
+        load_resource = " -R"
+
+    # we set the primary and secondary Lumina servers to 0.0.0.0 to disable Lumina,
+    # which sometimes provides bad names, including overwriting names from debug info.
+    #
+    # return values from open_database:
+    #   0 - Success (database not packed)
+    #   1 - Success (database was packed)
+    #   2 - User cancelled or 32-64 bit conversion failed
+    #   4 - Database initialization failed
+    #   -1 - Generic errors (database already open, auto-analysis failed, etc.)
+    #   -2 - User cancelled operation
+    ret = idapro.open_database(
+        str(path), run_auto_analysis=True, args=f"-Olumina:host=0.0.0.0 -Osecondary_lumina:host=0.0.0.0{load_resource}"
+    )
+    if ret not in (0, 1):
+        raise RuntimeError("failed to analyze input file")
+
+    logger.debug("idalib: waiting for analysis...")
+    ida_auto.auto_wait()
+    logger.debug("idalib: opened database.")
+
+    extractor = capa.features.extractors.ida.extractor.IdaFeatureExtractor()
+    fixup_idalib(path, extractor)
+    return extractor
+
+
+def fixup_idalib(path: Path, extractor):
+    """
+    IDA fixups to overcome differences between backends
+    """
+    import idaapi
+    import ida_funcs
+
+    def remove_library_id_flag(fva):
+        f = idaapi.get_func(fva)
+        f.flags &= ~ida_funcs.FUNC_LIB
+        ida_funcs.update_func(f)
+
+    if "kernel32-64" in path.name:
+        # remove (correct) library function id, so we can test x64 thunk
+        remove_library_id_flag(0x1800202B0)
+
+    if "al-khaser_x64" in path.name:
+        # remove (correct) library function id, so we can test x64 nested thunk
+        remove_library_id_flag(0x14004B4F0)
+
+
 @lru_cache(maxsize=1)
 def get_cape_extractor(path):
     from capa.helpers import load_json_from_path
@@ -227,33 +298,13 @@ def get_vmray_extractor(path):
     return VMRayExtractor.from_zipfile(path)
 
 
-GHIDRA_CACHE: dict[Path, tuple] = {}
-
-
+@lru_cache(maxsize=1)
 def get_ghidra_extractor(path: Path):
-    # we need to start PyGhidra before importing the extractor
-    # because the extractor imports Ghidra modules that are only available after PyGhidra is started
-    import pyghidra
-
-    if not pyghidra.started():
-        pyghidra.start()
-
-    import capa.features.extractors.ghidra.context
     import capa.features.extractors.ghidra.extractor
 
-    if path in GHIDRA_CACHE:
-        extractor, program, flat_api, monitor = GHIDRA_CACHE[path]
-        capa.features.extractors.ghidra.context.set_context(program, flat_api, monitor)
-        return extractor
+    extractor = capa.features.extractors.ghidra.extractor.GhidraFeatureExtractor()
+    setattr(extractor, "path", path.as_posix())
 
-    # We use a larger cache size to avoid re-opening the same file multiple times
-    # which is very slow with Ghidra.
-    extractor = capa.loader.get_extractor(
-        path, FORMAT_AUTO, OS_AUTO, capa.loader.BACKEND_GHIDRA, [], disable_progress=True
-    )
-
-    ctx = capa.features.extractors.ghidra.context.get_context()
-    GHIDRA_CACHE[path] = (extractor, ctx.program, ctx.flat_api, ctx.monitor)
     return extractor
 
 
@@ -914,20 +965,8 @@ FEATURE_PRESENCE_TESTS = sorted(
         ("mimikatz", "function=0x4556E5", capa.features.insn.API("advapi32.LsaQueryInformationPolicy"), False),
         ("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),
@@ -1015,20 +1054,20 @@ FEATURE_PRESENCE_TESTS = sorted(
         ("pma16-01", "file", OS(OS_WINDOWS), True),
         ("pma16-01", "file", OS(OS_LINUX), False),
         ("mimikatz", "file", OS(OS_WINDOWS), True),
-        ("pma16-01", "function=0x404356", OS(OS_WINDOWS), True),
-        ("pma16-01", "function=0x404356,bb=0x4043B9", OS(OS_WINDOWS), True),
+        ("pma16-01", "function=0x401100", OS(OS_WINDOWS), True),
+        ("pma16-01", "function=0x401100,bb=0x401130", OS(OS_WINDOWS), True),
         ("mimikatz", "function=0x40105D", OS(OS_WINDOWS), True),
         ("pma16-01", "file", Arch(ARCH_I386), True),
         ("pma16-01", "file", Arch(ARCH_AMD64), False),
         ("mimikatz", "file", Arch(ARCH_I386), True),
-        ("pma16-01", "function=0x404356", Arch(ARCH_I386), True),
-        ("pma16-01", "function=0x404356,bb=0x4043B9", Arch(ARCH_I386), True),
+        ("pma16-01", "function=0x401100", Arch(ARCH_I386), True),
+        ("pma16-01", "function=0x401100,bb=0x401130", Arch(ARCH_I386), True),
         ("mimikatz", "function=0x40105D", Arch(ARCH_I386), True),
         ("pma16-01", "file", Format(FORMAT_PE), True),
         ("pma16-01", "file", Format(FORMAT_ELF), False),
         ("mimikatz", "file", Format(FORMAT_PE), True),
         # format is also a global feature
-        ("pma16-01", "function=0x404356", Format(FORMAT_PE), True),
+        ("pma16-01", "function=0x401100", Format(FORMAT_PE), True),
         ("mimikatz", "function=0x456BB9", Format(FORMAT_PE), True),
         # elf support
         ("7351f.elf", "file", OS(OS_LINUX), True),

tests/test_ghidra_features.py

@@ -11,42 +11,95 @@
 # 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 importlib.util
+
+"""
+Must invoke this script from within the Ghidra Runtime Environment
+"""
+import sys
+import logging
+from pathlib import Path
 
 import pytest
-import fixtures
 
-import capa.features.common
-
-ghidra_present = importlib.util.find_spec("pyghidra") is not None and "GHIDRA_INSTALL_DIR" in os.environ
+try:
+    sys.path.append(str(Path(__file__).parent))
+    import fixtures
+finally:
+    sys.path.pop()
 
 
-@pytest.mark.skipif(ghidra_present is False, reason="PyGhidra not installed")
-@fixtures.parametrize(
-    "sample,scope,feature,expected",
-    [
-        (
-            pytest.param(
-                *t,
-                marks=pytest.mark.xfail(
-                    reason="specific to Vivisect and basic blocks do not align with Ghidra's analysis"
-                ),
-            )
-            if t[0] == "294b8d..." and t[2] == capa.features.common.String("\r\n\x00:ht")
-            else t
-        )
-        for t in fixtures.FEATURE_PRESENCE_TESTS
-    ],
-    indirect=["sample", "scope"],
-)
+logger = logging.getLogger("test_ghidra_features")
+
+ghidra_present: bool = False
+try:
+    import ghidra  # noqa: F401
+
+    ghidra_present = True
+except ImportError:
+    pass
+
+
+def standardize_posix_str(psx_str):
+    """fixture test passes the PosixPath to the test data
+
+    params: psx_str - PosixPath() to the test data
+    return: string that matches test-id sample name
+    """
+
+    if "Practical Malware Analysis Lab" in str(psx_str):
+        # <PosixPath>/'Practical Malware Analysis Lab 16-01.exe_' -> 'pma16-01'
+        wanted_str = "pma" + str(psx_str).split("/")[-1][len("Practical Malware Analysis Lab ") : -5]
+    else:
+        # <PosixPath>/mimikatz.exe_ -> mimikatz
+        wanted_str = str(psx_str).split("/")[-1][:-5]
+
+    if "_" in wanted_str:
+        # al-khaser_x86 -> al-khaser x86
+        wanted_str = wanted_str.replace("_", " ")
+
+    return wanted_str
+
+
+def check_input_file(wanted):
+    """check that test is running on the loaded sample
+
+    params: wanted - PosixPath() passed from test arg
+    """
+
+    import capa.ghidra.helpers as ghidra_helpers
+
+    found = ghidra_helpers.get_file_md5()
+    sample_name = standardize_posix_str(wanted)
+
+    if not found.startswith(fixtures.get_sample_md5_by_name(sample_name)):
+        raise RuntimeError(f"please run the tests against sample with MD5: `{found}`")
+
+
+@pytest.mark.skipif(ghidra_present is False, reason="Ghidra tests must be ran within Ghidra")
+@fixtures.parametrize("sample,scope,feature,expected", fixtures.FEATURE_PRESENCE_TESTS, indirect=["sample", "scope"])
 def test_ghidra_features(sample, scope, feature, expected):
+    try:
+        check_input_file(sample)
+    except RuntimeError:
+        pytest.skip(reason="Test must be ran against sample loaded in Ghidra")
+
     fixtures.do_test_feature_presence(fixtures.get_ghidra_extractor, sample, scope, feature, expected)
 
 
-@pytest.mark.skipif(ghidra_present is False, reason="PyGhidra not installed")
+@pytest.mark.skipif(ghidra_present is False, reason="Ghidra tests must be ran within Ghidra")
 @fixtures.parametrize(
     "sample,scope,feature,expected", fixtures.FEATURE_COUNT_TESTS_GHIDRA, indirect=["sample", "scope"]
 )
 def test_ghidra_feature_counts(sample, scope, feature, expected):
+    try:
+        check_input_file(sample)
+    except RuntimeError:
+        pytest.skip(reason="Test must be ran against sample loaded in Ghidra")
+
     fixtures.do_test_feature_count(fixtures.get_ghidra_extractor, sample, scope, feature, expected)
+
+
+if __name__ == "__main__":
+    # No support for faulthandler module in Ghidrathon, see:
+    # https://github.com/mandiant/Ghidrathon/issues/70
+    sys.exit(pytest.main(["--pyargs", "-p no:faulthandler", "test_ghidra_features"]))

tests/test_idalib_features.py

@@ -0,0 +1,63 @@
+# Copyright 2020 Google LLC
+#
+# 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
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# 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 pytest
+import fixtures
+
+import capa.features.extractors.ida.idalib
+
+logger = logging.getLogger(__name__)
+
+idalib_present = capa.features.extractors.ida.idalib.has_idalib()
+if idalib_present:
+    try:
+        import idapro  # noqa: F401 [imported but unused]
+    except ImportError:
+        idalib_present = False
+
+
+@pytest.mark.skipif(idalib_present is False, reason="Skip idalib tests if the idalib Python API is not installed")
+@fixtures.parametrize(
+    "sample,scope,feature,expected",
+    fixtures.FEATURE_PRESENCE_TESTS + fixtures.FEATURE_SYMTAB_FUNC_TESTS,
+    indirect=["sample", "scope"],
+)
+def test_idalib_features(sample, scope, feature, expected):
+    try:
+        fixtures.do_test_feature_presence(fixtures.get_idalib_extractor, sample, scope, feature, expected)
+    finally:
+        logger.debug("closing database...")
+        import idapro
+
+        idapro.close_database(save=False)
+        logger.debug("closed database.")
+
+
+@pytest.mark.skipif(idalib_present is False, reason="Skip idalib tests if the idalib Python API is not installed")
+@fixtures.parametrize(
+    "sample,scope,feature,expected",
+    fixtures.FEATURE_COUNT_TESTS,
+    indirect=["sample", "scope"],
+)
+def test_idalib_feature_counts(sample, scope, feature, expected):
+    try:
+        fixtures.do_test_feature_count(fixtures.get_idalib_extractor, sample, scope, feature, expected)
+    finally:
+        logger.debug("closing database...")
+        import idapro
+
+        idapro.close_database(save=False)
+        logger.debug("closed database.")