The Agent Skills Directory

[SAFE]: The skill uses standard shell utilities such as grep, glob, date, and head to perform local code analysis and documentation mining, which aligns with its primary research purpose.
[SAFE]: The skill utilizes platform-specific MCP tools (mcp__claude-flow__memory_usage, mcp__claude-flow__memory_search) for internal state management and inter-agent coordination. These are expected behaviors for an agent operating within a coordinated swarm environment.
[SAFE]: The use of user-controlled input (the $TASK variable) in the hooks section for storing research context is a legitimate functional requirement and does not present an exfiltration risk as it is stored locally within the environment's memory coordination system.

agent-researcher