sdr-ramp-cert
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input during its interactive training modes, which could potentially be used to influence the agent's internal logic or tool use.\n
- Ingestion points: The skill accepts user text during
quiz,roleplay, andcertmodes as defined inSKILL.md.\n - Boundary markers: The instructions lack explicit delimiters or instruction wrappers to differentiate user-provided dialogue from the agent's core instructions.\n
- Capability inventory: The skill has access to search tools for the Epiphan product catalog and knowledge base, and it possesses file-writing capabilities to generate HTML scorecards in the
outputs/directory.\n - Sanitization: There is no evidence of input validation or sanitization to prevent adversarial instructions from being processed during role-plays or scoring.
Audit Metadata