cost-discovery-coach-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external product data and possessing file system write capabilities.
- Ingestion points: The agent retrieves product specifications and facts from Epiphan Knowledge search and product catalog tools (SKILL.md).
- Boundary markers: No specific delimiters or instructions to ignore embedded malicious content within search results are defined, although the skill implements a fallback mechanism for unverified specs (SKILL.md).
- Capability inventory: The skill is configured to write and append outcome logs to local files within the ~/.claude/ directory (SKILL.md).
- Sanitization: A validation gate (check_my_copy) is employed to review customer-facing strings before delivery (SKILL.md).
Audit Metadata