What problem does it solve?
Enables precise routing of development tasks to external CLI-based AI coding agents based on their strengths, reducing manual triage and improving output quality and turnaround time for reviews, implementations, and bulk operations.
Core Features & Use Cases
- Agent Decision Matrix: Maps task types to primary and fallback agents (codex, qwen, gemini) for code review, implementation, refactoring, bulk changes, and documentation.
- Agent Capabilities & Workflows: Documents strengths/weaknesses of each agent and provides multi-step workflows for new feature development, large refactors, and documentation generation.
- Execution Helper: Includes a scripts/delegate.sh helper to run non-interactive or interactive agent commands and manage model selection and dry-run options.
- Templates & Examples: Provides task templates for reviews, implementations, and bulk operations to standardize delegation and review cycles.
Quick Start
Run the scripts/delegate.sh helper with an agent name and a task file to delegate the task non-interactively and inspect the returned output for review.