Archive task

Archive Task

Goal: move a task out of the active working set by relocating its folder tasks/TASK-<ID>/ → tasks/archive/TASK-<ID>/, preserving every file and the full history. This keeps tasks/ focused on live work and stops finished/abandoned tasks from cluttering "current task" scans — without deleting anything.

archive-task is a lifecycle skill, not a stage: it changes only the task's location (and logs why), never its stage or artifacts:. It's reversible — resume-task <ID> finds an archived task and restores it.

When to use

• A task is done and you want it off the active list.
• A task is being abandoned/superseded and shouldn't show up as live work.
• The user asks to archive / shelve / put away a task.
• Not for tasks you'll resume soon (use pause-task — it stays in place) and never to delete history (archiving moves, never removes).

Shared task state

Part of the task pipeline — see ../WORKFLOW.md → "Task lifecycle". This skill writes tasks/ (it moves a folder); archived ids are retired and never reused for new tasks.

Method

Input: an optional task id argument in any form (archive-task TASK-20260723-fix-login, archive-task fix-login, archive-task 007), normalized to the on-disk folder name TASK-<ID>: a full TASK-… id as written; a bare all-digit argument maps to a legacy numeric folder matched as written (don't re-pad or strip leading zeros — 7 and 007 both resolve to TASK-007 if that's the folder); anything else (a slug fragment, or a ticket key like PROJ-123) resolves against tasks/TASK-* and tasks/archive/TASK-* folder names — an exact match (TASK-<argument>) wins, else a unique substring match; several matches → list candidates and ask; none → say so and stop — never fall back to auto-picking another task. A match already under tasks/archive/ needs nothing: report that it's archived and stop. No argument → resolve the current task (below).

1. Locate the task: the normalized id argument, one named in conversation, else the most-recently-updated: task under tasks/TASK-* (skip tasks/archive/*, and don't auto-pick a status: paused task — to archive a paused one, name it explicitly). If ambiguous, list candidates and ask — don't guess. If the folder doesn't exist, say so; don't create anything.
2. Confirm intent — this moves a folder. State which task and its current stage/status, then:
   • If status: done → proceed (still confirm once). But first check the draft shipped: if review.md holds a commit/PR draft, look for a matching commit (git log) — if none exists, the work may be finished but unshipped; surface the draft to the user before archiving so it doesn't get buried in archive/.
   • If not done (active / blocked / paused — unfinished work) → ask the user to confirm archiving unfinished work, and capture a reason (e.g. "superseded by TASK-20260722-parser-rewrite"). Never archive live work silently.
3. Log before moving — get the real time (date "+%Y-%m-%d %H:%M %Z"), then in tasks/TASK-<ID>/task.md: bump updated: and append a dated Pipeline Log line - <YYYY-MM-DD HH:MM +TZ> archived: <reason> — carrying your agent label (format: ../WORKFLOW.md → Agent handoff). Leave stage and artifacts: untouched.
4. Move the folder to tasks/archive/TASK-<ID>/ (create tasks/archive/ if missing). Move the whole folder intact — every artifact comes along. Don't git rm/git mv or commit unless the user asks.
5. Confirm to the user: the task is archived at tasks/archive/TASK-<ID>/, and /resume-task <ID> restores it if needed.

When done

Report the task id, where it moved, and the reason. The task no longer appears in active scans; nothing was deleted.