debugging-and-error-recovery
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a collection of educational and procedural documentation. It guides the agent through structured debugging (SKILL.md), accessibility auditing (accessibility-checklist.md), performance tuning (performance-checklist.md), and security reviews (security-checklist.md).
- [SAFE]: No malicious command execution or unauthorized data access patterns were detected. The commands mentioned (e.g.,
npm test,git bisect,npx lighthouse) are standard development tools used for diagnostic purposes. - [INDIRECT_PROMPT_INJECTION]: The skill recognizes the risk of indirect prompt injection. In
SKILL.md, under 'Treating Error Output as Untrusted Data', it explicitly instructs the agent to treat error messages and logs as untrusted data and forbids the automatic execution of commands or navigation to URLs found within such data without user confirmation. - [SECURITY_HARDENING]: The
security-checklist.mdfile provides extensive best practices for preventing common vulnerabilities such as SQL injection, SSRF, and prompt injection, and includes a detailed section on managing the supply chain risk of dependency install scripts.
Audit Metadata