The Agent Skills Directory

[COMMAND_EXECUTION]: Orchestrates multi-agent communication by executing amq CLI commands such as send, watch, and drain to manage the collaborative workflow phases.\n- [DATA_EXFILTRATION]: Transmits user-provided design problems and agent-generated research findings to partner agent handles via the amq messaging system. While this is the intended function for collaborative design, it involves the transfer of codebase findings and technical analysis across agent boundaries.\n- [SAFE]: Enforces a critical human-in-the-loop safety checkpoint (Phase 5) where the proposed plan must be presented to and approved by the user before any execution or implementation actions are taken.

amq-spec