self-reflection
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- No Code (SAFE): The provided files consist of documentation and configuration examples. No executable scripts, binaries, or source code were included in the skill package for analysis.\n- Indirect Prompt Injection (SAFE): The skill reads from a local memory file (~/workspace/memory/self-review.md). This ingestion point is used for reviewing past reflections, which is the primary purpose of the skill. Because the content is intended to be agent-generated and no high-risk capabilities are exposed, the risk is minimal.\n
- Ingestion points: The
readcommand accesses the reflection log file.\n - Boundary markers: None specified in the documentation.\n
- Capability inventory: Commands are limited to logging and reading text-based reflections.\n
- Sanitization: Not explicitly mentioned, but the data source is local.
Audit Metadata