blender-forge
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill drives external tools including Blender and FFmpeg to perform 3D rendering and video encoding.
- Evidence: In
scripts/encode_turntable.py, therunfunction usessubprocess.run()with command lists. This implementation is secure as it avoids theshell=Trueparameter, preventing command injection vulnerabilities. - [INDIRECT_PROMPT_INJECTION]: The skill handles user-provided text for 3D inscriptions, which represents a potential injection surface.
- Ingestion points: Untrusted data enters the context through arguments like
--signs,--line, and--textinscripts/make_vessel.pyandscripts/make_tablet.py. - Boundary markers: Instructions are logically separated by script arguments; however, no explicit prompt delimiters are used for the text content itself.
- Capability inventory: The skill can write files (GLB, JSON, PNG) and execute local binaries (FFmpeg) via subprocess calls.
- Sanitization: Includes input validation such as profile checks in
scripts/make_vessel.pyand Unicode glyph resolution inscripts/signs.py. - [EXTERNAL_DOWNLOADS]: Mentions external dependencies required for operation.
- Evidence:
SKILL.mdidentifies Blender and FFmpeg as requirements and suggests installation via Homebrew. These are well-known and trusted software packages.
Audit Metadata