vmray: record command line info (#2515)

* vmray: record command line info

CHANGELOG.md

@@ -6,6 +6,7 @@
 
 - allow call as valid subscope for call scoped rules @mr-tz
 - support loading and analyzing a Binary Ninja database #2496 @xusheng6
+- vmray: record process command line details @mr-tz
 
 ### Breaking Changes
 

capa/features/extractors/vmray/__init__.py

@@ -35,6 +35,8 @@ class VMRayMonitorProcess:
     ppid: int  # parent process ID assigned by OS
     monitor_id: int  # unique ID assigned to process by VMRay
     image_name: str
+    filename: str
+    cmd_line: str
 
 
 class VMRayAnalysis:
@@ -160,7 +162,12 @@ class VMRayAnalysis:
                 self.sv2.processes[process.ref_parent_process.path[1]].os_pid if process.ref_parent_process else 0
             )
             self.monitor_processes[process.monitor_id] = VMRayMonitorProcess(
-                process.os_pid, ppid, process.monitor_id, process.image_name
+                process.os_pid,
+                ppid,
+                process.monitor_id,
+                process.image_name,
+                process.filename,
+                process.cmd_line,
             )
 
         # not all processes are recorded in SummaryV2.json, get missing data from flog.xml, see #2394
@@ -170,6 +177,8 @@ class VMRayAnalysis:
                 monitor_process.os_parent_pid,
                 monitor_process.process_id,
                 monitor_process.image_name,
+                monitor_process.filename,
+                monitor_process.cmd_line,
             )
 
             if monitor_process.process_id not in self.monitor_processes:

capa/features/extractors/vmray/extractor.py

@@ -86,7 +86,7 @@ class VMRayExtractor(DynamicFeatureExtractor):
 
     def get_process_name(self, ph) -> str:
         monitor_process: VMRayMonitorProcess = ph.inner
-        return monitor_process.image_name
+        return f"{monitor_process.image_name} ({monitor_process.cmd_line})"
 
     def get_threads(self, ph: ProcessHandle) -> Iterator[ThreadHandle]:
         for monitor_thread_id in self.analysis.monitor_threads_by_monitor_process[ph.inner.monitor_id]:

capa/features/extractors/vmray/models.py

@@ -136,11 +136,20 @@ class FunctionReturn(BaseModel):
     from_addr: HexInt = Field(alias="from")
 
 
+def sanitize_string(value: str) -> str:
+    # e.g. "cmd_line": "\"C:\\Users\\38lTTV5Kii\\Desktop\\filename.exe\" ",
+    return value.replace("\\\\", "\\").strip(' "')
+
+
+# unify representation
+SanitizedString = Annotated[str, BeforeValidator(sanitize_string)]
+
+
 class MonitorProcess(BaseModel):
     ts: HexInt
     process_id: int
     image_name: str
-    filename: str
+    filename: SanitizedString
     # page_root: HexInt
     os_pid: HexInt
     # os_integrity_level: HexInt
@@ -148,7 +157,7 @@ class MonitorProcess(BaseModel):
     monitor_reason: str
     parent_id: int
     os_parent_pid: HexInt
-    # cmd_line: str
+    cmd_line: SanitizedString
     # cur_dir: str
     # os_username: str
     # bitness: int
@@ -306,8 +315,9 @@ class Process(BaseModel):
     monitor_id: int
     # monitor_reason: str
     os_pid: int
-    filename: str
+    filename: SanitizedString
     image_name: str
+    cmd_line: SanitizedString
     ref_parent_process: Optional[GenericReference] = None