know-me
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill manages long-term memory via local file system operations in a specific directory structure (~/.claude/projects/). This is a standard implementation for maintaining state and does not involve unauthorized file access or persistence mechanisms targeting system configuration.
- [SAFE]: Privacy and security rules are explicitly defined in the instructions, prohibiting the storage of secrets, API keys, or sensitive personal data. This mitigates the risk of credential exposure through the memory system.
- [PROMPT_INJECTION]: The skill maintains an indirect prompt injection surface as it processes untrusted user data and stores it for later context injection.
- Ingestion points: Persistent markdown files in
~/.claude/projects/<project-path>/memory/read by the agent at the start of sessions. - Boundary markers: None explicitly defined in the instructions for separating stored memory from new instructions.
- Capability inventory: File reading and writing to the project's memory directory.
- Sanitization: None specified; the skill relies on the agent's internal safety guidelines and instructions to exclude sensitive data.
Audit Metadata