Repository navigation
docs: keep posted warm brief messages as plain history - #20979
Conversation
Posted preparing, result and failure messages stay as ordinary history; a stale result is dropped unposted. Removes the retraction and persist-before-surface clauses and records the English fixed lines.
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — 🟡 CONCERNSDesign-level review of Design-Verdict: CONCERNS A stale attempt now leaves “Preparing your current-task summary. You can wait here.” in history for good, and no answer ever follows it. Watch
[DESIGN-REVIEWED] f0db087 Description read[DESCRIPTION-READ] 7e0f41262758997b7f132f661ece257047f6c0908134225eca3b3ca2996fef58 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
chenmingwei23
left a comment
There was a problem hiding this comment.
Approving: readiness: passed, all lanes green, no outstanding change requests.
Problem / Motivation
Goal: Make the warm task brief RFC match the simple plain-message shape now implemented in #20422.
#20778 described stale generated messages being retracted by row id and every row being withheld until its disk write succeeded. The user chose simpler behavior: a posted crewmate message is ordinary history and stays.
Why it matters
The implementation must read one accepted, non-conflicting design from
main. Without this amendment, review correctly blocks #20422 for not retracting rows the RFC still says it retracts.Not a goal
What changed (motivation → approach → change)
docs/request-for-change/rfc-delayed-task-brief.md:assistantis required by the ordinary message path, and the reader filters cover metadata-stripping recent projections and session transfer. The heading of 3.4 now reads "Preparation before the model" and undo describes the two-release rollback.Backwards compatibility
Compatible: this changes design text only and no running code.
Removes nothing: no product code or accepted trigger is removed.
Tests
Manual verification
Reviewed sections 3.3, 3.4, 3.5, 3.7, user-visible outcomes, validation, rollout, undo, and the decision record together.
Screenshots / video
N/A — documentation-only design amendment with no rendered product change.
Related Issues
Part of #20156
Related implementation: #20422
Checklist
Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license.