task-orchestration

Task orchestration in a repository is managed by driving a single authoritative task list through a strict plan, execute, and iterate loop tracked in a .copilot-todo.yaml file. This provides a single source of truth to reduce context switching and keep work auditable.

To seed a YAML task list from a Markdown plan, copy the plan template to plan.md, edit it with Markdown checkboxes, and run the task orchestrator script with the --plan-md and --plan-mode regen init arguments to generate the YAML state.

Yes, workflow management enforces per-task commits by requiring a clean working state before advancing and recording updates and commits for each individual task. This maintains repository integrity and ensures single-task focus with traceability.

YAML-based task tracking is operated through a set of CLI commands including init, status, next, update, add, and peek. These commands support hands-off automation or interactive use for driving the plan-driven workflow.

Yes, the task orchestration workflow requires the PyYAML dependency to handle YAML-based state management. You also need Python 3 to run the orchestrator scripts that drive the plan-to-YAML conversion and task execution loop.

The best way to maintain repository task status for software teams is using a plan-driven loop that tracks status and ordering directly in a YAML file. This enforces single-task execution and incremental commits across the project.