Executes assigned agentic tasks with full context retrieval and lifecycle management via the Granary CLI.
The Granary Task Executor skill transforms Claude into a specialized sub-agent capable of completing specific programming or administrative tasks assigned by an orchestrator. By leveraging the Granary CLI context hub, the skill retrieves precise implementation details, file paths, and project steering conventions, ensuring that work follows established patterns without requiring extensive manual exploration. It handles the entire task lifecycle—from initial context building and implementation to progress reporting and completion—while maintaining safety in multi-agent environments through task leasing and conflict detection.
Key Features
01Complete task lifecycle management including start, done, and block states
02Direct integration with project-specific steering files and architectural patterns
03Automated context retrieval using granary handoff for precise task execution
043 GitHub stars
05Rich progress reporting with custom comment kinds like progress and decisions
06Multi-agent safety features with task leasing and heartbeat monitoring
Use Cases
01Implementing specific feature modules based on pre-researched architectural plans
02Collaborating in multi-agent environments where tasks are distributed and claimed asynchronously
03Executing targeted bug fixes across multiple files identified by a planning agent
