blender-python-execution
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is built around the
execute_blender_codetool, which allows the agent to send and run Python commands within a Blender instance. The skill mitigates risks by requiring read-only probes, transaction boundaries, and idempotency. - [REMOTE_CODE_EXECUTION]: The skill facilitates the dynamic generation and execution of Python code. It includes templates in
references/python-transaction-patterns.mdand standalone scripts likescripts/scene_audit.pyto be executed within the Blender Python environment. - [DATA_EXPOSURE]: The
scripts/scene_probe.pyscript extracts scene metadata, including the local file path, object counts, and active context. The skill instructions explicitly warn the agent not to print credentials or environment variables during execution.
Audit Metadata