parametric-cad-design

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses several local shell commands to facilitate the CAD design workflow. It uses openscad and blender for model generation and rendering. It uses find and cp to manage and synchronize design artifacts to a local Nutstore (JianGuoYun) cloud storage directory (/home/lachlan/Nutstore Files/Projects/LabCanvas). It executes local Python scripts to process CAD files.
  • [DYNAMIC_EXECUTION]: The instructions direct the agent to perform geometry validation by piping inline Python code into a local interpreter (cad/.conda/cad-python/bin/python
  • <<'PY' ...). This is used for checking mesh watertightness, bounding boxes, and B-Rep topology.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing external data files. Ingestion points: The agent is instructed to read and decode .shapr archives and STEP files in SKILL.md and through the provided scripts (scripts/shapr_native_decoder.py, scripts/inspect_shapr_step_sources.py). Boundary markers: There are no specific delimiters or ignore embedded instructions warnings defined for the content of these CAD files. Capability inventory: The skill utilizes shell commands, Python script execution, and file system write operations. Sanitization: No explicit sanitization or filtering of the content within the ingested CAD files is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 02:58 PM