diagnose

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a logical framework for software debugging and performance regression analysis. All strategies (e.g., unit testing, bisection, diffing) follow industry standard practices for developers.
  • [COMMAND_EXECUTION]: The skill includes a shell script (scripts/hitl-loop.template.sh) designed for Human-in-the-Loop (HITL) debugging. The script uses standard read commands to interact with a user during a local debugging session and prints captured output for the agent to analyze. There is no evidence of command injection or unsafe execution from untrusted sources.
  • [EXTERNAL_DOWNLOADS]: The skill references standard web tools like Playwright and Puppeteer for UI testing, but does not provide URLs to download untrusted binaries or execute remote scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 08:46 AM
Security Audit — agent-trust-hub — diagnose