agent-workflow

Multi-agent workflow coordination manages parallel agent work through a five-phase lifecycle: REQUIREMENTS, ARCHITECTURE, IMPLEMENTATION, VERIFICATION, and REFLECTION. This Skill spawns and tracks Tier-1 and Tier-2 agents using a Dotagent-based pattern, synchronizing their outputs across phases so work progresses collectively rather than in isolation.

Workflow orchestration enforces structured phases: REQUIREMENTS (define scope), ARCHITECTURE (design structure), IMPLEMENTATION (build code), VERIFICATION (test and validate), and REFLECTION (analyze outcomes). This Skill manages transitions between phases and ensures agents complete each stage before moving forward, reducing coordination overhead.

Spawn subagents using Dotagent-based patterns within a two-tier hierarchy: Tier-1 agents handle high-level coordination, Tier-2 agents execute specialized tasks. This Skill automates spawning, assignment, and lifecycle management so you initialize once and let the workflow distribute work across the agent tiers.

Yes, this Skill is applied at project inception to initialize the multi-agent workflow. It establishes the phase structure, agent hierarchy, and spawning rules upfront so every new development task follows consistent orchestration patterns from start to finish.

Phase transitions are managed automatically by the workflow. When agents complete their assigned phase work, the coordinator validates outputs, synchronizes state, and triggers the next phase. This prevents bottlenecks and ensures downstream agents receive validated inputs.

Yes, the tiered agent hierarchy and phase-gated structure scale to complex projects. Tier-1 agents coordinate high-level phases while Tier-2 agents parallelize implementation tasks, so workflow remains manageable as complexity grows.