blender-motion-state-inspection
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external Blender files (.blend), which act as untrusted data input. Ingestion points: The agent interacts with .blend files provided by the user using command-line tools. Boundary markers: The skill lacks explicit instructions or markers to distinguish between data and potentially malicious instructions that could be embedded in the scene's metadata or object names. Capability inventory: The agent is granted Bash, Write, and Edit tools, allowing it to perform shell command execution and filesystem modifications. Sanitization: There are no defined mechanisms for sanitizing the output of the Blender scripts or the contents of the scene files before they are processed by the agent.
- [DYNAMIC_EXECUTION]: The inspection workflow involves the agent generating a Python script (e.g., collect_motion_state.py) and executing it via the Blender interpreter (blender --python). This represents runtime script generation and execution for automation purposes.
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to invoke the blender executable with various arguments to perform background processing of 3D assets.
Audit Metadata