Repository navigation
docs: record that a bot's sessions tab opens a session - #36187
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Documentation only. The new decision entry follows the supersede process in Notes
[DESIGN-REVIEWED] 76438a0 Description read[DESCRIPTION-READ] d229657c8463fea258bd3e0fea1cf21a5f7118085a12d4fc8b451d588390adf8 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 |
Opus 5.5 Review — ✅ human override acceptedReviewed Human judgment by @buluoray overrides the Opus 5.5 finding for [OPUS-OVERRIDE] 76438a0 Verdict recorded from an authorized human decision for commit False positive or not applicable? A repository writer can comment: |
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: |
Code change approver recommendation: Request changesWhy: the decider found 1 problem and estimates an 85% chance that this change introduces a defect, at or above the 80% level where it asks for changes. This docs-only PR adds a decision entry that partly supersedes the 2026-11-03 Sessions-tab entry, plus its row in the README index. The content checks out against the repo. The RFC sentences it cites exist ("no longer opens it" and "no report yet" in rfc-mate-sessions-folder-or-lineage.md, lines 100 and 104). So does the quote from the 2026-11-06 ledger entry. The named maintainer, Ray Xu (@buluoray), is in MAINTAINERS.md, and no existing entry is edited. The problem is that the entry is not signed at this head: its key Evidence line reads PENDING. Entries are immutable once merged, so landing this head would permanently record a decision without the on-record maintainer restatement that rule 4 requires. The PR description says as much. Swap PENDING for the comment link, then land it with the maintainer override. Problems found:
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 |
|
I decided this, as maintainer: On a Bot's Sessions tab a click on a row still expands it in place into a recap. The opened recap carries one button that opens that session in the chat view ("Reopen session" for a closed session, which is reopened from its saved conversation first); everything else in the recap stays read-only. A session whose work item has no report, or that has no work item, shows a short excerpt of the latest message the Bot sent to that worker, through the same redaction as the row, in place of "No report yet". "No report yet" stays only for a session the Bot never sent a message. This supersedes, in part, |
7e8b22c to
1d03218
Compare
|
I decided this, as maintainer. It replaces my earlier comment on this pull request (#36187 (comment)), which said "short excerpt" and left out the RFC sentences. On a Bot's Sessions tab a click on a row still expands it in place into a recap. The opened recap carries one button that opens that session in the chat view ("Reopen session" for a closed session, which is reopened from its saved conversation first). It is the recap's one action; everything else in the recap stays read-only. A session whose work item has no report, or that has no work item, shows a bounded excerpt of the latest message the Bot sent to that worker, in place of "No report yet": a few lines, through the same redaction as the session's row, with the rest reached through the button. "No report yet" stays only for a session the Bot never sent a message. This supersedes, in part, |
1d03218 to
e7779ee
Compare
|
/ai-review override gpt e7779ee owner-decision: The maintainer decided this superseding entry and restated it on this PR: #36187 (comment). The rule blocks a Supersedes entry by design; this override is the approval it asks for. |
|
/ai-review override fable e7779ee owner-decision: The maintainer decided this superseding entry and restated it on this PR: #36187 (comment). The rule blocks a Supersedes entry by design; this override is the approval it asks for. |
Human judgment recorded@buluoray recorded a owner-decision release (class
This decision applies only to this commit. A new push requires a new judgment. |
Human judgment recorded@buluoray recorded a owner-decision release (class
This decision applies only to this commit. A new push requires a new judgment. |
A Bot's Sessions tab expands a row in place into a recap and gives no way to open the session, and a session with no bound work item always reads "No report yet". This entry records the change of mind: the opened recap gets a button that opens the session, and a session with no report shows a bounded excerpt of the latest message the Bot sent to that worker. It supersedes, in part, the 2026-11-03 Sessions tab entry. The evidence line for the maintainer's restatement is filled in before this merges. Part of #36175 Part of #36177
e7779ee to
76438a0
Compare
|
/ai-review override fable 76438a0 reissue: Same ruling as #36187 (comment) for this rebased head; the entry is unchanged and the maintainer's decision is #36187 (comment). |
Human judgment recorded@buluoray recorded a reissue release (class
This decision applies only to this commit. A new push requires a new judgment. |
bolichen97
left a comment
There was a problem hiding this comment.
Read the whole diff at 76438a0: one new decision entry and its README row.
- The entry is narrow. A row click still expands the recap in place, and the recap gains one action, open or reopen in the chat view, while everything else stays read-only.
- The "No report yet" replacement is bounded. It's a few lines of the Bot's latest message to that worker, through the same redaction as the session's row. "No report yet" stays for a session the Bot never sent a message.
- The supersession is scoped. It names the exact clause of
2026-11-03-a-crewmates-sessions-leave-the-sidebar-for-its-sessions-tab.mdand the two RFC sentences it replaces, says every other rule stands, and links the issues (#36175, #36177) and #28931. - The README row matches the entry. This is a docs-only change with no code and no regex.
All checks are green. Approving. Auto-merge is left as you set it.
Problem / Motivation
Goal: Record the maintainer's decision that a Bot's Sessions tab can open a session from its recap, and that a session with no report shows the task the Bot last gave it.
A Bot's Sessions tab lists the worker sessions it started. A click expands a row in place into a recap. Two things about that recap were decided on 2026-11-03 and are now changed:
Both behaviours follow a recorded decision, so no code may change them until a new entry supersedes it. This pull request is that entry and nothing else.
Why it matters
Without the entry, the two fixes are regressions by this repository's own rule (
docs/decisions/README.md, rule 6) and cannot land. With it, a person can open a worker session from the one place that lists it, and the recap stops reading as empty for sessions that hold real work.Not a goal
What changed (motivation → approach → change)
One new decision entry and its row in the entries table.
docs/decisions/2026-11-12-a-bots-sessions-tab-opens-a-session-and-shows-the-task-it-was-last-given.md2026-11-03-a-crewmates-sessions-leave-the-sidebar-for-its-sessions-tab.md.README.mdWhat the entry says:
The maintainer signed this entry. Ray Xu (@buluoray) restated the decision in a comment on this pull request, as
docs/decisions/README.mdrule 4 asks, and the entry's first Evidence line links that comment. A chat message is not evidence, so the entry quotes none.An agent wrote the entry's text. The decision is the maintainer's.
The reviewer rule blocks this pull request by design.
docs/decisions/README.mdrule 3: a superseding entry "ships in its own pull request: therecorded-decisions-are-not-regressionsrule blocks that PR by design, and a maintainer's override is what lets it land". No override is requested here; that is the maintainer's step.Backwards compatibility
Compatible: documentation only. The 2026-11-03 entry stays byte-identical (
scripts/check_decisions_history.pypasses), and no code reads this file.Removes nothing: no entry is edited, renamed or deleted.
Tests
No test changes: the diff is two documentation files.
Checks run on this head:
DECISIONS_BASE_REF=origin/main python3 scripts/check_decisions_history.py: every entry at the base is intact at the head.bash scripts/docs-lint.sh: all documentation checks passed.scripts/check_brand_name.pyandscripts/check_comment_history.pyagainst the merge base: clean.Manual verification
N/A: no behaviour changes. The entry was read against the format in
docs/decisions/README.mdand against the entry and RFC section it changes.Related Issues
Part of #36175
Part of #36177
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.