What problem does it solve? Keeping a Markdown-based task system consistent is hard: task IDs collide between parallel agents, ledgers drift out of sync with task files, and malformed metadata slips into commits. This Skill defines the canonical task and inbox file formats and provides the tooling that reserves IDs atomically, regenerates derived ledgers, and validates the whole docs/tasks_manager/ tree. ## Core Features & Use Cases - Atomic ID reservation: at reserve inbox <slug> and at reserve task <PREFIX> <TYPE> <slug> create placeholder files under a lock so concurrent agents never claim the same ID. - Ledger sync and validation: at ledger sync regenerates _active.md, _done.md, and generated area pages; at ledger check fails on duplicate IDs, malformed metadata, status-directory mismatches, and stale generated files. - Execution support: at task brief <TASK-ID> --phase N writes a self-contained implementer brief per phase, and at task run-state init|check manages the resume map for phased execution. - Use Case: After triaging an inbox idea into task AUTH-001, you implement it phase by phase, then run at ledger check to confirm every task file, roadmap reference, and generated ledger is consistent before committing. ## Quick Start Ask the agent to reserve a new task ID for the auth area and validate the task ledger with at ledger check after editing any task file.
