What problem does it solve? Coordinating multiple AI workers across tickets, dependencies, and merges is error-prone when done manually. Gru acts as an autonomous leader that dispatches Minion workers, verifies their claims, manages integration locks, and recovers from failures without constant operator supervision. ## Core Features & Use Cases - Autonomous worker dispatch: Launch and manage Minion workers through Hermod with concurrency, attempt, and recovery limits. - Dependency-aware coordination: Track durable assignments and dependency readiness in a shared SQLite state store, integrating one change at a time under a merge lock. - Bounded recovery and lifecycle control: Resume interrupted runs, reconnect existing workers, stop engagements cleanly, and escalate concrete blockers with evidence. - Use Case: You have five approved tickets from a gated spec. Run Gru with the objective and ticket list, and it dispatches workers, verifies each PR through the required review gate, merges them in dependency order, and closes tickets only after verified completion. ## Quick Start Ask Gru to start leading work on your approved tickets by running $gru start with your objective and ticket IDs.
