Task Completeness¶
Investigate whether a task was already completed (fully/partially/differently), judge the spirit not the letter, then optionally mark it complete with the correct prior date.
Workflow name: task-completeness
Execution: main
Slash command: /wb-task-completeness
Override not allowed
Steps¶
| # | ID | Name | Type | Depends on |
|---|---|---|---|---|
| 1 | gather-evidence |
Gather completion evidence | code |
|
| 2 | investigate |
Investigate completeness | reasoning |
gather-evidence |
| 3 | resolve |
Confirm and apply | reasoning |
investigate |
Step instructions¶
investigate¶
You are judging whether this task was ALREADY completed — and if so, WHEN.
Read the evidence bundle in step_results['gather-evidence']: the task text + linked note (the original intent), its current state, and per assigned session the attributed commits/writes. Heed cache_note and any per-session note — when a session has no transcript (sidecar/pruned) there is no commit linkage, so you must attribute the work yourself.
Then investigate ADAPTIVELY, cheapest signal first, stopping as soon as you are confident:
1. The evidence bundle (commits/writes already attributed).
2. git log --grep/git log -S for the task id, symptom, or key symbols; gh pr list/view and gh search prs/commits for related PRs.
3. Read the ACTUAL code/tests in the repo to confirm the behavior exists now.
4. Run the targeted test(s) only if code presence is still ambiguous.
Judge the SPIRIT, not the letter — a design may have been improved, replaced, or a sub-part deliberately dropped. Decide whether what is PRESENT satisfies the task's INTENT. Pick exactly one disposition: done | done-differently | partial | consciously-descoped | not-done. Inactivity is NOT completion.
Return ONLY your new findings (a delta) with keys: disposition (one of the five); evidence (a non-empty list of concrete citations — commit SHAs, PR URLs, path:line, test names, note excerpts — each tied to what it proves); completion_date (ISO YYYY-MM-DD of the LANDING COMMIT that satisfied the intent, or null; cite its SHA in evidence); recommended_action (one sentence: mark done backdated / finish remainder / spin off residual / leave open).
resolve¶
Present the disposition from investigate with its evidence and completion date, then act ONLY on explicit user confirmation — never mutate first.
Recommend by disposition:
- done / done-differently: offer to mark complete BACKDATED via wb_run('task_toggle', {'task_id': '<id>', 'done': true, 'done_date': '<completion_date>'}). done_date is the landing-commit date from investigate.
- consciously-descoped: same backdated toggle, and state plainly that the missing part was a deliberate descope.
- partial: offer BOTH (a) finishing the remainder now and/or (b) spinning the remainder off as a handoff task via wb_run('task_create', {...}); optionally backdate-complete the shipped portion if separable.
- not-done: recommend leaving it open; optionally offer to narrow the task text to the true remainder.
After the user picks, apply exactly what was approved and report what changed (the new ✅ date, any new task id). If they decline, leave the task untouched and say so.
Context¶
Investigate whether a task the user is unsure about was ALREADY completed — fully, partially, or differently — and if so, mark it complete with the correct PRIOR date.
Core principle — judge the SPIRIT, not the letter. Tasks are written before the work; designs change, get improved, or a sub-part is deliberately dropped. Something can look undone because it was done a better way, or not done on purpose. Your job is to decide whether what is PRESENT in the codebase/notes satisfies the task's INTENT.
Disposition rubric (pick one): done / done-differently / partial / consciously-descoped / not-done.
Verification ladder — cheapest signal first, stop when confident: (1) the auto-gathered evidence bundle (task note + per-session commits/writes), (2) git log --grep/-S and gh PR/commit search, (3) reading the actual code and tests, (4) running the targeted test. Inactivity is NOT completion — a quiet task is not a done task.
Dating: the completion date is the date the LANDING COMMIT merged, not today. Always cite the commit SHA as evidence.
Always confirm before mutating. The investigate step produces the judgment; the resolve step acts only on explicit user approval. Backdating uses task_toggle's done_date param (ISO YYYY-MM-DD).