What problem does it solve? When an organization runs multiple Squad instances across separate repositories, agents lack a standard way to discover each other's capabilities, share context, and hand off work without creating tight coupling or leaking internal state. ## Core Features & Use Cases - Squad Discovery: Locate other squads by reading .squad/manifest.json files from well-known paths, upstream configs, or a central registry, revealing each squad's capabilities and accepted work types. - Work Handoff Protocol: Delegate tasks by creating labeled GitHub issues in the target squad's repository with context, acceptance criteria, and links back to the originating issue. - Feedback Loop Tracking: Poll delegated issues with gh issue view, update the originating issue with status changes, and close the loop when work merges. - Use Case: A frontend squad needs a Prometheus metrics endpoint owned by the platform squad. The agent checks the platform squad's manifest, confirms it accepts issues, creates a [cross-squad] labeled issue with acceptance criteria, and tracks it to completion. ## Quick Start Ask the agent to discover available squads and delegate a specific task to the squad that owns the required capability.
