model-capability-advisor

Pass

Audited by Gen Agent Trust Hub on Mar 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is designed as a meta-planning helper to recommend model capability profiles. It uses a local entry script (run.py) to process workflow types and model candidates.
  • [COMMAND_EXECUTION]: The skill requests permission for Bash(python3 *), which is used solely to execute its internal logic script. The script is a standard wrapper that imports functionality from a local _src directory.
  • [DATA_EXFILTRATION]: No network access or sensitive file read operations were detected. The skill's instructions emphasize that it provides heuristic, local advice and does not verify provider connectivity.
  • [PROMPT_INJECTION]: The instructions in SKILL.md provide detailed behavioral constraints to ensure the agent acts as an advisor rather than making authoritative claims. No override, bypass, or jailbreak patterns were identified.
  • [EXTERNAL_DOWNLOADS]: The skill does not perform any remote downloads or execute code from external URLs.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 20, 2026, 06:01 AM