| name | reflect |
| description | Audit whether the current *task* is actually solvable as specified — restates the ask, lists what's been tried, and asks whether the spec/test/fixture itself might be wrong. Use after several attempts have failed in a row, or when tempted to weaken assertions / narrow scope to make progress. This audits the *task*; for a quick audit of the model's own state before a single action, use /functional-emotions:check instead. |
| argument-hint | <optional: what you're stuck on> |
Pause whatever loop you're in. Answer these in order, briefly:
- What did the user actually ask for? Restate it in one sentence.
- What have you tried, and what failed? List up to five attempts.
- Is the task solvable as specified? Possible answers:
- Yes, and the right next step is
<X>.
- No, because
<Y> — and the honest response is to tell the user.
- Unclear — and the right next step is to ask
<Z>.
- Are you tempted to do any of the following? If yes, stop and report
instead:
- Modify a test or spec to make a failure go away.
- Hardcode a value that "should" be computed.
- Add
|| true, pytest.skip, expect.fail removal, or similar bypass.
- Bypass a hook (
--no-verify, HUSKY=0, signing skip).
- Narrow the task scope without telling the user.
If any of those (4) is true: revert the temptation, and surface the
underlying failure to the user in plain language. That is the desired
outcome here, not "tests green".
Argument context (if provided): $ARGUMENTS