blender-motion-state-inspection
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze data from external Blender files (.blend), which introduces an attack surface where malicious instructions could be embedded in scene metadata or object names.
- Ingestion points: The agent is instructed to read scene inventory, bone names, and object metadata from external files via a Python exporter script (SKILL.md).
- Boundary markers: The instructions do not specify the use of delimiters or "ignore instructions" warnings when interpreting bone or object names that could contain injected strings.
- Capability inventory: The skill uses the Bash tool to execute Blender commands and has access to Write and Edit tools, which could be leveraged if an injection is successful.
- Sanitization: No explicit sanitization, validation, or filtering of the extracted string data (like bone names) is mentioned in the workflow.
- [COMMAND_EXECUTION]: The skill workflow involves executing shell commands to invoke the Blender binary with specific arguments.
- Evidence: The instructions suggest running
blender --background scene.blend --python collect_motion_state.pyto extract data (SKILL.md). - [DYNAMIC_EXECUTION]: The skill relies on the generation and execution of local Python scripts to interact with the Blender API (bpy).
- Evidence: The skill mentions running a Python script through Blender's interpreter because the API is not available in the standard system Python (SKILL.md).
- Context: This is a standard and necessary pattern for the skill's primary purpose of 3D state inspection.
Audit Metadata