Repository navigation
feat(dashboard): the warm welcome reports like a manager: ask, progress, what needs you - #23847
Conversation
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: the progress count comes only from the gateway's ledger, briefs stored in the older shape still draw, and no reader outside the patch reads Notes
[DESIGN-REVIEWED] 5649df0 Description read[DESCRIPTION-READ] 0bdd732de4f78f464811f943d50b99ed19c245d72113728cf00c2a15df1a4f93 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 |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Nothing blocks. Notes
What this change shipsInventory (8 items) — 5 justifiedIntent: when the owner comes back, show them status, what needs them, real progress and their original ask. ADDITION.
[FIRST-PRINCIPLES-REVIEWED] 5649df0 Description read[DESCRIPTION-READ] 6cc26c797edf3d725027fbb92a09e114c144a16769fa93357e3933634737ccee That sha256 names the PR description this verdict read. Recompute it with |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing blocks. Every element this PR adds appears in a screenshot, and the blind reader read the status chip, headline and progress bar correctly. The card has no clickable controls, as the PR states. Notes
[UX-REVIEWED] 5649df0 Description read[DESCRIPTION-READ] 8756fca3d3ee759638b70043ca1b84c10c835a51170023cb119b5d8b0b29f5d5 That sha256 names everything this verdict read: the PR description, plus the 10 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 — ✅ 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: |
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: |
|
2c415ad to
fd7d659
Compare
|
Rebased: main merged #23828 into src/kiro_crew/dashboard/warm_task_brief.py and test/test_warm_task_brief.py, which made this PR CONFLICTING (fd7d659 -> f19d75a on 0dab198). Resolved by keeping both: the snapshot carries #23828's running_workers and this PR's progress. test_warm_task_brief.py + test_slot_running_readers.py 355 passed; warmTaskBrief/ChatPane.preparing vitest 69 passed; flake8/isort/mypy/black clean. CI re-running. |
fd7d659 to
f19d75a
Compare
|
f19d75a to
0507f66
Compare
|
Bundle Size Gate red here is inherited from main: pristine main @ 0dab198 already builds ModelEffortDropdown over its 500 KB budget (505.0 KB per #23867). This PR's merge ref measures 507.9 KB, so the card adds about 2.9 KB. #23867 re-measures the ceiling to 531 KB, which covers that. Waiting for #23867 to land, then this PR rebases and re-runs the gate; no override requested. |
|
Rebased: the fresh run after close/reopen still measured ModelEffortDropdown against a 500 KB budget because GitHub's merge ref (a1967ed) was still built on 0dab198, before #23867's 531 KB ceiling (3503760). Rebased 0507f66 -> bcbfa5d onto main, which carries #23867. Warm-brief pytest 353 passed, warmTaskBrief vitest 66 passed. No code change. CI re-running. |
0507f66 to
bcbfa5d
Compare
…ss, what needs you
|
bcbfa5d to
5649df0
Compare
buluoray
left a comment
There was a problem hiding this comment.
Approving at 5649df0ab9fa68b85e5ad468e7b0bee441bbe166.
The returning owner now gets a status, what needs them, a real progress count and what is left, instead of four thin fields. The count comes from the work ledger, and the model's numbers are ignored (_goal_progress, and _parse_result takes progress from the snapshot). Where the current goal's items cannot be named (unreadable or legacy header, item count different from created_total, unparseable or tied creation times), the card shows no count rather than a wrong one.
Checked:
- Every new field, including
extratitles, goes through_clean(redaction, marker defang, cap). The card draws each one as a plain text node, and the renderer re-validatesstatusandprogresson its own. - All 8 new card keys are present in
en.manual.json,en-XAand all 12 translated catalogs, with both placeholders kept inprogress_count. The backend label rows arezip(strict=True), so a missing label fails at import. - No
docs/decisions/entry covers the warm brief, so nothing is contradicted. The change is recorded as a dated amendment on the acceptedrfc-delayed-task-brief.md, and the session/history specs and feature map are updated in the same commit. - Locally on this head: backend 354 passed,
warmTaskBrief.test.tsx66 passed,src/i18n1035 passed,lint:i18nOK,tsc -p tsconfig.app.jsonclean. - 10 mutations, each with a named red test: gateway progress dropped, count-mismatch guard, tie guard, unreadable header, status coercion, empty
extrasections,extratitle skipping_clean, and on the frontend the progress bound, thenowfallback and status validation. None survived.
Non-blocking follow-up: in Hindi, Bengali and Russian the card uses the informal you while the copied plain text (_TEXT_LABEL_ROWS) uses the formal one, so Copy reads in a different register from the card. This predates the PR for the existing labels, but it now covers more lines.
Bring the delayed-task-brief RFC and its directory index into agreement with the document's amendments and with what shipped on main. The warm task brief shipped in #20422 and its 2026-10-18, 2026-10-22 and 2026-10-24 amendments shipped in #21750, #22828, #22839 and #23847, but the RFC still read `accepted` ("nothing built yet") with implementation-prs [] and a 2026-10-14 audit stamp. The front matter now records status implemented, those five PRs, and last-audited 2026-10-26 at 1cf77b5; the section 1 status and "Audited on" lines match it; and the snapshot's third field is renamed "Suggested actions" to match the section 3.5 output it feeds. The README index row no longer describes the superseded preparing-status design or says implementation waits for the RFC. It reads implemented, names the shipped PRs, and summarises the shipped trigger and card. The audit paragraph gains a dated re-audit sentence for the row. Closes #20887
Bring the delayed-task-brief RFC and its directory index into agreement with the document's amendments and with what shipped on main. The warm task brief shipped in #20422 and its amendments shipped in #21750, #22828, #22839, #23847, #24512 and #24307, but the RFC still read `accepted` ("nothing built yet") with implementation-prs [] and a 2026-10-14 audit stamp. The front matter now records status implemented, those seven PRs, and last-audited 2026-10-28 at 7cbf872; the section 1 status and "Audited on" lines match it; and the snapshot's third field is renamed "Suggested actions" to match the section 3.5 output it feeds. The README index row no longer describes the superseded preparing-status design or says implementation waits for the RFC. It reads implemented, names the shipped PRs, and summarises the shipped trigger, the free-form brief card and the leave-after-opener rule. The audit paragraph gains a dated re-audit sentence for the row. Closes #20887
Bring the delayed-task-brief RFC and its directory index into agreement with the document's amendments and with what shipped on main. The warm task brief shipped in #20422 and its amendments shipped in #21750, #22828, #22839, #23847, #24512 and #24307, but the RFC still read `accepted` ("nothing built yet") with implementation-prs [] and a 2026-10-14 audit stamp. The 2026-10-28 writer amendment (section 3.8, #25433) is design only: main has no wake turn and no `dashboard.warm_brief_wake` setting. So the front matter now records status partial, those seven PRs, and last-audited 2026-10-29 at 4d36177; the section 1 status line names what shipped and that section 3.8 is not built; the "Audited on" line matches; and the snapshot's third field is renamed "Suggested actions" to match the section 3.5 output it feeds. The README index row no longer describes the superseded preparing-status design or says implementation waits for the RFC. It reads partial, names the shipped PRs, summarises the shipped trigger, the free-form brief card and the leave-after-opener rule, and names the unbuilt writer amendment. The audit paragraph gains a dated re-audit sentence for the row. Closes #20887
…21712) Bring the delayed-task-brief RFC and its directory index into agreement with the document's amendments and with what shipped on main. The warm task brief shipped in #20422 and its amendments shipped in #21750, #22828, #22839, #23847, #24512, #24307, #25881 (Crew Mode scope), #26500 (fixed progress orb) and #26652 (first part of the 2026-10-31 redesign), yet the front matter and README row still read `accepted` ("nothing built yet") with implementation-prs [] and a 2026-10-14 audit stamp. Not built on main: the 2026-10-28 writer amendment (section 3.8; main has no wake turn and no `dashboard.warm_brief_wake` setting), the 2026-11-01 trigger amendment (#26938 is open), and the live checks, links and copied option buttons of section 3.9.3. So the front matter now records status partial, those ten PRs, and last-audited 2026-11-01 at fed53fc; the section 1 status line names what shipped and what is not built; the "Audited on" line matches; and the snapshot's third field is renamed "Suggested actions" to match the section 3.5 output it feeds. The README index row no longer describes the superseded preparing-status design or says implementation waits for the RFC. It reads partial, names the shipped PRs, summarises the shipped trigger, the one-message brief with its fixed card and quiet return, and names the unbuilt parts. The audit paragraph gains a dated re-audit sentence for the row. Closes #20887
Problem / Motivation
Goal: When the owner comes back to a crewmate chat, the warm welcome reads like a manager's report: is it on track, what must the owner do, how far is it from the goal (counted from the real work ledger), and what did they ask for.
The warm welcome card (#22839) shows four short fields. The owner says it carries too little to act on. It does not say whether the work is on track, how much is done, or what is left, and people forget what a task line was about.
Why it matters
The brief is the first thing the owner reads after time away. If it cannot tell them where things stand and what they must do, they have to scroll the transcript or ask, which is the work the brief exists to save.
Not a goal
MateWelcomeCard,mateGreeting,member_recap).What changed (motivation → approach → change)
nowheadlineN of M donebar, counted by the gateway over the current goal's work items; the model's numbers are ignoredremaining, one line from the modelextrasections{title, items}for a risk, a cost or a choiceThe gateway side lives in
dashboard/warm_task_brief.py; the card ischat/WarmBriefCard.tsx.The prompt now casts the model as a professional manager reporting to the owner who hired it: lead with on-track and what needs the owner, remind them what they asked, say how far from the goal in real terms, no filler, never an invented number. The snapshot it reads carries the same
accepted/totalcount, so its "what is left" line agrees with the bar. An unknown status falls back to "waiting on you" when something needs the owner, else "on track"; "on track" or "done" above a Needs you box reads as "waiting on you".Every new field goes through the same redaction, control-marker defang and caps as before, and the card draws each one as a plain text node. The plain-text
contentcopy follows the card's order, with labels in the request's language. The new card labels are added to every locale (locales/en.manual.jsonand the translated catalogs).The decision is recorded in
docs/request-for-change/rfc-delayed-task-brief.md(a dated 2026-10-24 amendment in the status list and section 5.1, on top of the accepted RFC). The session and history specs and the feature map name the new fields;website/scripts/capture-warm-brief-card.mjscaptures the new card.Backwards compatibility
Compatible: a stored brief in the older shape (
asked,now,needs_you,next) still draws,nowas the headline and with no status chip; a malformed one keeps the ordinary bubble, as before. A model reply in the oldernowor prosemessageshape is read as the headline.Removes nothing: the older fields are still read; only new fields are written.
Tests
test/test_warm_task_brief.py: the full return-to-result path stores the new shape and its plain text in report order; every language has every label.test/test_warm_task_brief_shape.py(the pure parse/cap/redact tests, split out to stay under the 2,000-line test-file cap): progress comes from the ledger even when the model returns other numbers, counts only the current goal's items after a new goal, and is left out with no items or when the goal's items cannot be told apart; status vocabulary fallback; caps and item limits on every field and list, includingextra; malformedextrasections dropped; redaction, rejoined tokens and every control marker defanged in every new field,extratitles and items included; prose-only and headline-less replies refused; Chinese labels.website/src/test/warmTaskBrief.test.tsx: card section order; empty sections left out; no bar without a count; older shape renders; status and progress validation; caps; markup in every field stays text; chat search counts match the card's text nodes.Nearby inputs checked: no work items (tested), all items done (tested,
2 of 2), model omits headline (tested, refused), prose only (tested, refused → failure line), very long fields (tested, capped),[OPTIONS:insideextra(tested, defanged).Proof: not provable -- other: a feature change, not a fix; the new tests pin new behaviour
Manual verification
Captured the real Members page with
node scripts/capture-warm-brief-card.mjs(dark and light, 1280 px and 320 px for before/after): card drawn, no link/button/img inside, bar only with a count, reload draws the same card, card fits the viewport, no page errors.Screenshots / video
After (dark, 1280 px):
After (light, 320 px) and a blocked brief in Chinese:
Older brief, no work items, all done, injection
Related Issues
Closes #23827
Checklist