orchestratelisted

# orchestrate Use for larger delivery flows and multi-agent coordination. ## Use When - The task is large enough that parallel work will save time or add confidence. - Independent workstreams can be defined with clear inputs, outputs, and ownership. - One lead agent can own synthesis, integration, and final verification. - Work spans multiple repos, skill chains, or requires worktree separation. ## Do Not Use When - The task is small, tightly coupled, or faster to complete in one session. - Multiple agents would fight over the same files or the same mutable context. - The user wants one narrow implementation rather than orchestration. ## Core Responsibilities - Decide whether work should be serial, parallel, or worktree-split - Choose the controlling plan - Keep one source of truth for current state - Ensure each branch of work converges cleanly - Plan and run multi-agent work without losing integration quality ## Inputs / Prereqs - The goal, success criteria, and integration owner. - Candidate workstreams and the dependencies between them. - The quality gates or review checkpoints that must run after synthesis. - References on team topologies, coordination patterns, or example team boards (if available). ## Workflow 1. Decide whether parallelism buys time, confidence, or separation of concerns. - Done when: serial/parallel/worktree-split decision is recorded and communicated to team. 2. Split the work into independent tracks with an owner, expected output, and