guideline-generation
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements strong privacy and security constraints, requiring the agent to redact PII, anonymize company names, and flag sensitive information during the synthesis process.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from various external sources which could serve as a vector for indirect prompt injection.
- Ingestion points: Brand documents (PDF, PPTX, DOCX, MD, TXT) and conversation transcripts from services like Gong, Granola, and Notion.
- Boundary markers: The skill does not define specific string delimiters for inputs, but it mandates a 'Quality Check' step using a quality-assurance agent to verify no PII is exposed.
- Capability inventory: The skill has the capability to write and rename files within the
.claude/directory of the user's working folder for saving and archiving guidelines. - Sanitization: The instructions explicitly require redacting customer names and contact information and anonymizing excerpts before output.
Audit Metadata