canonical-team-lead

You can coordinate multiple AI coding agents by using a lead agent to orchestrate task creation, assignment, and execution via predefined CLI bridge commands. This ensures structured multi-agent development workflows without manual coordination.

A lead AI agent maintains team state by enforcing state persistence in TEAM.md, AGENTS.md, and TASKS.md files. These files act as a single source of truth for tracking task creation, assignment, and execution across the team.

Canonical actions for task orchestration include creating, assigning, starting, accepting, reopening, and posting updates for tasks. These are executed through predefined CLI bridge commands to manage the workflow automatically.

Yes, you can orchestrate a team of a planner, coder, and reviewer automatically. A lead AI can coordinate these agents to deliver a feature from specification to code review without manual intervention.

The best way to start an AI coding team workflow is to ask the lead skill to create a new task titled with your feature name and assign it to the appropriate coder agent using the canonical CLI commands.