What problem does it solve?
Guides Team Leads in supervising and recovering delegated AI agents to prevent wasted work, contain scope drift, and accelerate reliable convergence during multi-agent execution.
Core Features & Use Cases
- Active Steering & Background Management: Monitor inline sessions, steer early, and manage autonomous background agents with partition constraints and resume semantics.
- Parallel Research & Result Merging: Spawn 3–5 focused research sub-agents, deduplicate and synthesize findings, and produce a compact merged context for downstream implementation.
- Health Checks, Circuit Breaker & Escalation: Detect stuck or looping agents, apply cadence-based checks, open circuits after repeated failures, and escalate via dead-letter queues or dispute records.
- Use Case: When introducing a cross-cutting feature, run parallel research for frontend, backend, and CMS ownership, then synthesize results and delegate implementation with strict file partitions.
Quick Start
Use the orchestration-protocols skill to run a health check across active agents, compact prior phase context, and apply escalation or circuit-breaker recovery actions.