debugging-and-error-recovery
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill provides structured guidelines for debugging, utilizing standard local tools like npm and git for testing and regression analysis.
- [INDIRECT_PROMPT_INJECTION]: The skill includes defensive instructions specifically targeting indirect prompt injection. Ingestion points: Error output, logs, and stack traces processed during debugging. Boundary markers: Explicit instructions to treat output as data to analyze rather than instructions to follow. Capability inventory: npm and git shell commands. Sanitization: Directs the agent to surface instructions found in logs to the user rather than executing them automatically.
- [COMMAND_EXECUTION]: The skill references common development commands such as npm test and git bisect. These are standard practices for the intended debugging context and do not represent malicious command execution.
Audit Metadata