agent-introspection-debugging
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides an instructional framework for error capture and diagnosis without executing code.
- [INDIRECT_PROMPT_INJECTION]: The workflow ingests external data such as error messages and tool history into the agent context in SKILL.md. Boundary markers are provided via Markdown headers (e.g., ## Failure Capture). The capability inventory includes environment auditing and workspace analysis through referenced ECC skills. No explicit sanitization or filtering of ingested error logs is mentioned, representing a standard injection surface that is necessary for the skill's primary debugging purpose.
Audit Metadata