engineering-workflow-guide
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a documentation and routing guide for managing complex engineering tasks. It does not contain executable scripts, network operations, or sensitive file system access.
- [PROMPT_INJECTION]: No malicious prompt injection patterns were detected. The skill incorporates 'Anti-Rationalization Guardrails' and 'Anti-Hallucination Guardrails' specifically designed to enforce adherence to the curated skill catalog and prevent the agent from inventing non-existent capabilities.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user requests as untrusted input to perform intent classification. While this ingestion point exists, the skill's instructions focus on routing to predefined, packaged skills listed in a controlled catalog, effectively limiting the potential impact of adversarial input at this layer.
Audit Metadata