What problem does it solve?
Coordinating multiple AI agents on a complex feature is error-prone: agents claim success without proof, tasks overlap, and failures go unnoticed. This Skill automates multi-agent orchestration by decomposing work into priority-tiered tasks, spawning specialist CLI subagents in parallel, and enforcing verification before accepting results.
Core Features & Use Cases
- Automated Task Decomposition and Dispatch: Breaks a feature request into specialist tasks (backend, frontend, mobile, QA), classifies each into domain tags, and spawns agents via native runtime dispatch or the
oma agent:spawn fallback.
- Iterative Review Loop: Every completed agent passes mechanical self-checks,
oma verify, and QA cross-review, with structured feedback fed back on failure and retry limits enforced.
- Memory-Based Coordination: Uses session, task-board, progress, and result files with strict ownership rules so concurrent agents never conflict.
- Use Case: Ask to implement a full-stack JWT authentication feature; the orchestrator spawns backend, frontend, and QA agents in parallel, verifies each deliverable, retries failures, and compiles a final summary.
Quick Start
Ask the agent to orchestrate implementing a full-stack feature in parallel, for example: run the authentication feature across backend, frontend, and QA agents automatically.
