skill-creator

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • Subprocess Command Execution: The utility scripts include commands that orchestrate execution runs via system subprocess calls.
  • Why it is a concern: Spawning external system processes can sometimes be manipulated to execute arbitrary commands if input parameters are not safely handled.
  • Evidence: Python scripts scripts/run_eval.py, scripts/improve_description.py, and eval-viewer/generate_review.py leverage subprocess.Popen and subprocess.run to call the system's claude binary or execute environment-scoped testing hooks.
  • Context: This behavior is limited to programmatic automation for executing test prompts, measuring process timings, and initializing local data viewers under the direct workflow initiated by the developer.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 02:01 PM
Security Audit — agent-trust-hub — skill-creator