The Agent Skills Directory

[SAFE]: The skill implements a three-layer workflow (search, timeline, and fetch) for efficient data retrieval from the claude-mem persistent database. This approach optimizes token usage and follows logical data-handling practices.
[DATA_EXFILTRATION]: No external network operations or unauthorized data transfer patterns were identified. The skill interacts exclusively with defined MCP tools (search, timeline, get_observations) for internal data management.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection as it ingests historical session data. Ingestion point: get_observations tool (SKILL.md). Boundary markers: Absent. Capability inventory: No dangerous execution capabilities (subprocess, eval, or file-write) were identified across the provided file. Sanitization: Absent. No malicious behavior or behavior-override instructions were detected.

mem-search