parallel-coordinator

What problem does it solve?

This Skill efficiently orchestrates multiple specialized AI agents working concurrently or sequentially, ensuring optimal context sharing, preventing duplicate work, and maintaining high quality through medium-sized task batching and frequent check-ins.

Core Features & Use Cases

• Medium-Sized Batching: Breaks agent work into manageable 15-25 minute batches for balanced control and efficiency, allowing early issue detection.
• Context Partitioning & Synchronization: Manages isolated agent contexts, preventing duplicate research and synchronizing knowledge at critical points (e.g., API contract definitions).
• Session Checkpointing: Enables pause/resume capabilities for long-running or multi-day workflows, preserving progress and context across interruptions.
• Use Case: For a full-stack feature like "Real-time Notifications," this skill can launch backend and frontend agents in parallel, synchronize their work at the API contract definition, and allow you to pause and resume the entire coordination process across sessions, saving you time and ensuring consistency.

Quick Start

This skill is typically invoked by task-orchestrator for Level 2 tasks.

To manually trigger for a set of parallel tasks:

Invoke skill with:

feature_name: "Real-Time Notifications"

tasks: [

{"agent": "fastapi-backend-expert", "description": "Implement WebSocket server"},

{"agent": "react-frontend-architect", "description": "Implement Notification UI"}

]

session_dir: ".artifacts/real-time-notifications/{{timestamp}}"