blender-python-execution
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill's primary function is to generate and execute Python code within a Blender environment using the
bpyAPI via a Model Context Protocol (MCP) adapter. This is managed through a 'transaction loop' pattern (Inspect-Mutate-Verify) and provided templates inreferences/python-transaction-patterns.mdto ensure predictability and idempotency. - [INDIRECT_PROMPT_INJECTION]: The skill processes user requests to mutate Blender scenes, creating a surface where untrusted input could influence code generation. The skill mitigates this by providing dedicated audit and probe scripts (
scripts/scene_audit.py,scripts/scene_probe.py) to verify the scene state before and after execution, and by instructing the agent to use stable names and explicit object retrieval instead of relying on context-dependent indices. - [COMMAND_EXECUTION]: The skill provides tools to execute code inside the Blender process. The instructions explicitly forbid unsafe operations such as installing add-ons, enabling scripts from untrusted sources, or modifying system-level configurations, focusing strictly on scene data manipulation.
Audit Metadata