cognitive-testing
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions identify episodic memory and state databases as ingestion points for external data that could contain malicious instructions. It provides mitigation strategies to ensure system integrity.
- Ingestion points: SKILL.md (Steps 3 and 4) refers to the agent loading episodic memory and reading from a state database (state.db).
- Boundary markers: The instructions specify that episodic memory must be sanitized and that the system must detect and reject fabricated history.
- Capability inventory: The skill defines high-level testing procedures but does not include scripts or tools capable of file system writes, network requests, or subprocess execution.
- Sanitization: Explicit instructions are provided to sanitize episodic memories before storage and to use isolated mock environments for testing state-dependent logic.
Audit Metadata