using-blender
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted 3D asset files such as .blend, .fbx, and .glb which could contain malicious embedded scripts or configurations.
- Ingestion points: Workflow involves reading and inspecting .blend, .fbx, .obj, .glb, and glTF files as described in SKILL.md and SKILL.en.md.
- Boundary markers: The skill includes a "Security Rules" section explicitly instructing the agent to only execute audited or self-written scripts and to verify licenses for external assets.
- Capability inventory: The skill utilizes the blender CLI and the bpy (Blender Python API) for automation, which grants filesystem access.
- Sanitization: Instructions recommend limiting scope, network access, and directory permissions when using MCP servers or executing arbitrary Python code.
- [DYNAMIC_EXECUTION]: The workflow relies on the generation and execution of Python scripts using the Blender API (bpy) to automate asset processing.
- Evidence: SKILL.en.md directs the agent to "use a small bpy script that makes inputs, outputs, and errors explicit" for automation tasks.
- [COMMAND_EXECUTION]: The skill provides templates for executing Blender in headless mode via the system shell.
- Evidence: SKILL.md contains a PowerShell command template: blender.exe --background --factory-startup --python ".TOPICS.ROBLOX_tools\verify_asset_visual.py" -- --fbx "<asset.fbx>" --out "" --json.
Audit Metadata