skills/mitsuhiko/agent-stuff/openscad/Gen Agent Trust Hub

openscad

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION] (SAFE): The skill executes the openscad binary and local utility scripts to render and validate 3D models. This is the primary intended function and is implemented using safe shell practices (e.g., using arrays for arguments to prevent shell injection).
  • [DATA_EXPOSURE] (SAFE): File access is restricted to the models and configuration files provided by the user. No access to sensitive system credentials or private directories (~/.ssh, ~/.aws) was detected.
  • [PROMPT_INJECTION] (LOW): The skill has an indirect prompt injection surface as it parses external .scad and .json files. * Ingestion points: tools/extract-params.sh (reads .scad files) and tools/render-with-params.sh (reads .json files). * Boundary markers: Absent. * Capability inventory: File writing (STL/PNG) and local command execution (OpenSCAD). * Sanitization: Uses regex-based parsing via Python for parameter extraction.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:27 PM