parametric-part-cad

Warn

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The script scripts/build_part.py uses runpy.run_path to execute Python scripts provided as command-line arguments. This allows for the execution of arbitrary Python code contained within the target script file at runtime.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary workflow involves taking untrusted user input (natural language part descriptions) and transforming it into Python code which is then executed. This creates an attack surface where a malicious user could craft a description designed to trick the agent into generating and executing unintended code.
  • Ingestion points: User-provided part descriptions and dimensions are processed in SKILL.md workflow steps 1-3.
  • Boundary markers: The workflow uses a parameter table as an intermediary, but there are no explicit instructions to ignore embedded instructions in the user's description.
  • Capability inventory: The skill has the capability to write files to the local file system and execute arbitrary Python code via runpy in scripts/build_part.py.
  • Sanitization: While the skill performs geometry validation (sanity checks), it does not include sanitization or filtering to prevent code injection patterns during the script generation process.
  • [COMMAND_EXECUTION]: The skill's workflow explicitly instructs the agent to execute shell commands (python scripts/build_part.py ...) to build and export the CAD geometry.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 7, 2026, 06:02 AM