What problem does it solve? Coordinating multiple coding agents or AI developer tools on one task often leads to duplicate work, conflicting file edits, and lost source of truth. This Skill plans which agent owns each step — planning, implementing, reviewing, or testing — and defines safe handoffs so parallel work does not collide. ## Core Features & Use Cases - Agent Selection Guidance: Maps task steps to internal partner/gemini lanes first, with a comparison table for external tools like Claude Code, Codex, Gemini, Cursor, Amp, Droid, OpenCode, and Antigravity as fallback. - Execution Patterns: Provides single-agent, sequential, parallel, and review-loop patterns with explicit file ownership, merge strategy, and an arbiter checklist for final integration. - Handoff Templates: Ships copy-ready assignment, brief, review-request, and final-integration templates in references/handoff-templates.md. - Use Case: A large feature spans an API package and a web app. Use this Skill to split ownership between two agents, assign an integrator for shared files, and require test evidence before merging outputs. ## Quick Start Ask the assistant to plan a multi-agent workflow for your task, for example: orchestrate a feature implementation across two coding agents with a review loop and a final integrator.
