Skills
Navigate
skills/dawiddutoit/custom-claude/openscad-workshop-tools/Gen Agent Trust Hub

openscad-workshop-tools

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill instructions and referenced templates are limited to 3D geometry generation and local rendering. No malicious behaviors such as data exfiltration, obfuscation, or persistence were found.
  • [COMMAND_EXECUTION]: The skill utilizes the openscad command-line tool to render generated .scad files into PNG images for preview. This execution is scoped to local files and directories.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes tool dimensions and features from potentially untrusted external sources like manuals or product pages.
  • Ingestion points: SKILL.md (Step 1) specifies gathering dimensions and features from external manuals or product pages.
  • Boundary markers: The skill uses a standardized OpenSCAD code structure to isolate parameters from geometry logic.
  • Capability inventory: The agent writes .scad files to the workshop/tools/ directory and executes the openscad CLI tool.
  • Sanitization: Input is interpreted as numeric parameters or color arrays within the OpenSCAD language, which lacks system-level access capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 02:46 AM