mem-search
Pass
Audited by Gen Agent Trust Hub on Apr 11, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill retrieves untrusted historical data from the claude-mem database, creating an indirect prompt injection surface where malicious instructions from past sessions could influence current behavior.
- Ingestion points: Data enters the agent context via the search, timeline, and get_observations tools as specified in SKILL.md.
- Boundary markers: No delimiters or instructions to ignore embedded commands within the retrieved memory are present.
- Capability inventory: The skill is restricted to database queries and does not define dangerous operations like file system writes or network access.
- Sanitization: The instructions do not include mechanisms for sanitizing or validating the retrieved historical content.
Audit Metadata