multi-agent-coordinationlisted

# Multi-Agent Coordination A reusable system for running multiple Claude Code terminals on the same project safely. Extracted from production use (Etihad CRM), generalized for any stack. ## When to activate Activate this skill when the user: - Wants to run 2+ Claude Code sessions in parallel on one repo - Asks about multi-agent / parallel-Claude / squad workflows - Mentions file collisions between AI sessions - Asks for a shared task board / kanban for AI agents - Wants a planner+developer split (architect terminal vs coder terminals) - Asks how to do file locking, approval gates, or commit coordination across AI sessions ## The five core mechanics 1. **Terminal roles** — each session has an explicit role: Planner (label `P` — reviews, never writes code) or Developer (labels `T1`, `T2`, `T3`, … — implements). See `references/terminal-roles.md`. 2. **File locks** — `active_files.md` is a shared lock registry. Append a line before editing, remove after. Timestamped so stale locks expire. See `references/lock-protocol.md`. 3. **Shared kanban** — `active_tasks.md` has four sections (TODO / AWAITING REVIEW / BLOCKED / DONE). Single source of truth for what each terminal is doing. See `references/lock-protocol.md` (kanban section). 4. **Approval gate** — developers never commit until the planner has reviewed the uncommitted diff and explicitly approved. See `references/approval-gate.md`. 5. **Git workflow** — pick Variant A (per-task feature branches + PR) or Variant B (direct