pip-responder
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external documents (Performance Improvement Plans) and personal backstories provided by the user. \n
- Ingestion points: Processes user-provided PIP documents and narrative backstories in SKILL.md. \n
- Boundary markers: No explicit delimiters are defined to separate ingested text from instructions. \n
- Capability inventory: The skill generates text-based reports, plans, and email templates. It does not utilize system tools, file system writes, or network requests. \n
- Sanitization: No evidence of sanitization for ingested content.
Audit Metadata