debugging-and-error-recovery
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides purely instructional content for systematic debugging and error recovery, following industry-standard workflows for triage, reproduction, and verification.
- [SAFE]: Includes an explicit security section on 'Treating Error Output as Untrusted Data', which instructs the agent to treat error messages and logs as potentially untrusted input, serving as a mitigation against indirect prompt injection.
- [COMMAND_EXECUTION]: Contains examples of standard development commands such as 'npm test', 'git bisect', and 'npm run build'. These are provided as part of a guide for manual or agent-assisted debugging and are appropriate for the skill's stated purpose.
Audit Metadata