diff options
| author | Craig Jennings <c@cjennings.net> | 2026-08-20 11:53:07 -0700 |
|---|---|---|
| committer | Craig Jennings <c@cjennings.net> | 2026-08-20 11:53:07 -0700 |
| commit | 84b558ae0988ee0646c40a3f3020878fafbd2c63 (patch) | |
| tree | af07d207fa65ef3cfff86b1014371d486a28127b /claude-templates/.ai/workflows/drill-deck-review.org | |
| parent | 435f622da565ca8ccb34b44b8917186b30d03a7d (diff) | |
| download | rulesets-84b558ae0988ee0646c40a3f3020878fafbd2c63.tar.gz rulesets-84b558ae0988ee0646c40a3f3020878fafbd2c63.zip | |
refactor(inbox): drop deferral-by-rename from the disposition ladder
An item now gets filed or deleted. Nothing stays in inbox/ under a renamed prefix.
The retired rung renamed a deferred item to PROCESSED-<original> and told you not to let those pile up, without giving that instruction any enforcement. It failed twice. They accumulated, to 40 here before a manual sweep last month and 114 across the rest of the fleet today. And a task citing one as its source went dangling when that sweep ran, which is the opposite of what deferral was for.
Both come from the same defect. PROCESSED-* was a third state, neither a tracked task nor gone. It sat in inbox/ where inbox-status hid it and no review cycle owned it. Filing and deleting are the two states that already have homes.
I left the three PROCESSED-* exclusions in place and marked them transitional. Dropping them today would have made 114 files pending at once and blocked five projects behind a cleanup nobody asked for. A sweep task owns clearing them, and each comment points at it.
Phase C's third option was the deferral. It's now delete, which is the genuinely distinct alternative once filing is the recommendation.
Diffstat (limited to 'claude-templates/.ai/workflows/drill-deck-review.org')
0 files changed, 0 insertions, 0 deletions
