Repository navigation
docs(rfc): propose waiting-on-you cards that pin an agent's ask above the composer - #39294
unstablebrainiac wants to merge 1 commit into
Conversation
986354e to
701cae1
Compare
First Principles Review (Opus 5.5, fork) — ⏭️ skippedRevision |
UX Review (Opus 5.5, fork) — ⏭️ skippedRevision |
Design Review (Opus 5.5, fork) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: this PR only adds a draft RFC and its index row. It changes no code, and it leaves the 2026-09-13 decision amendment for a maintainer to rule on. Notes
[DESIGN-REVIEWED] 8f6432f Description read[DESCRIPTION-READ] d0b4d65dd83932ea70831dfa3c64f8843d817146a6045b20ee55116340664caa 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 (fork) — ✅ no blocking findingsReviewed Review detailsNo findings. |
Opus 5.5 Review (fork) — ✅ no blocking findingsReviewed |
701cae1 to
05efde2
Compare
|
Phase 1 of this design is up for review as #39362. It follows this RFC, with three differences listed in its description (a paused crew queues an answer; edit-and-resend clears asks like rewind; Shared Crew teammates get 403 on the answer and dismiss routes). This RFC will be updated to match once those are reviewed. Phase 2 (loop guidance) waits for acceptance here. |
05efde2 to
8f6432f
Compare
Code change approver recommendation: A person should decideWhy: the decider estimates a 7% chance that this change introduces a defect, within the 12% limit for approving on its own, but it is held at tier 2: design-doc-gate: D. This PR is docs only: a new draft RFC and one index row in docs/request-for-change/README.md. It changes no code, state, limits or runtime behaviour. The RFC does not edit the 2026-09-13 Tier: 2, held for a person: design-doc-gate: D. An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit |
|
This changes behaviour users already rely on, so it needs a design sign-off from a maintainer before it can merge. A maintainer will post the decision here. To propose a design, follow the request-for-change process in CONTRIBUTING.md (docs/request-for-change/). Review result on this head: the diff is docs only (the RFC and one index row), so no code runs. PR Readiness is green. We found no blocking issue. We checked the code symbols the RFC cites against this head and current main, and they exist where the RFC says. The RFC also handles the 2026-09-13 Optional notes for the revision before acceptance:
|
8f6432f to
4261fae
Compare
Problem / Motivation
Goal: Propose, as an RFC for a maintainer to accept, a card that an agent pins above the composer to say what it needs from the person, kept there until the person answers it or writes, the agent clears it, or it expires.
When an agent is blocked on the person, only the turn that first said so explains the ask. In a watch loop, every later cycle adds another "still waiting on you" reply on top of it, so the person scrolls back through the chat to find what is needed. Chips from
[OPTIONS:]come only from the newest reply, so a blocked loop must restate them every cycle to keep them on screen.Why it matters
Loops that wait on a person for hours or days produce one visible reply per cycle that says nothing new, and the actual ask scrolls out of view. Non-loop chats lose the ask too, as soon as the person asks a side question. #20961 asks for the same "waiting on you" signal for a plain-text ask.
Not a goal
ask_question, native question cards,[OPTIONS:]chips orneeds_input.What changed (motivation → approach → change)
A new RFC,
docs/request-for-change/rfc-waiting-on-you-cards.md, with statusdraft, and its row indocs/request-for-change/README.md.waiting_on_youpins one ask (a key, a title, optional detail, 0 to 6 options);waiting_on_you_clearremoves it. At most 4 per session.change_cards.json, so it survives restarts and lazy tab restore. The transcript gets a display-onlycardrow.ask_idand the stored options, and is idempotent.waitingkind in the needs-me card for answering and dismissing from the sidebar.It amends one line of the 2026-09-13
[OPTIONS:]decision, the one that tells a blocked loop to restate its options, and it asks a maintainer for that ruling explicitly. Independent reviews by Opus 5.5 and GPT 5.6 Sol ran over five rounds; their findings are folded in.Backwards compatibility
Compatible: the PR adds documentation only.
Removes nothing: no code, key or document is changed apart from one added index row.
Tests
Proof: not provable -- docs or config only
Manual verification
N/A -- docs only. Every code citation in the RFC was checked against
640338acb1by the author and by two reviewers.Real-machine verification
N/A -- the diff adds Markdown documents; nothing in it runs.
Related Issues
Part of #20961
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...); PR Hygiene does not count commitsContribution License Agreement
Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.