Skip to content

feat(chat): write the warm welcome as a plain four-part recap in the gateway - #33805

Merged
buluoray merged 1 commit into
mainfrom
feat/warm-welcome-four-part
Nov 13, 2026
Merged

buluoray merged 1 commit into
mainfrom
feat/warm-welcome-four-part

Conversation

@iamwhatever

@iamwhatever iamwhatever commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: On return, the warm welcome is a short, plain recap in the owner's language that states only facts gateway code read, and explains what the owner must decide in words.

Owner-directed shape. This shape is the owner's own ruling of 2026-11-09. It replaces the closed #33032, whose cards, live hiding and model framing the owner did not take. The owner directed ONE pull request holding both the RFC change (docs/request-for-change/rfc-delayed-task-brief.md, new dated section 3.11) and the code. Design and First Principles: the RFC change here is the owner's ruling, not an agent's proposal.

A truth audit of 71 welcomes (2026-11-02 to 2026-11-07) found the gateway's own pull-request rows right 34 of 34 times and the model's sentences false 11 of 58 times. The owner asked a follow-up question in 32 of 55 replies.

Why it matters

The welcome is the first thing an owner reads on return. When one model sentence in five is false, the owner re-checks everything and the welcome costs time instead of saving it.

Not a goal

  • Cards, pictures, widgets, live hiding, change counters, or any model text in the welcome.
  • Card buttons: main's copied [OPTIONS:] chip row stays the click path, unchanged.
  • A Board refresh on conductor wakes, a timer, or any network read on Board page load.
  • Goal-board changes: every other Board path stays as it is.

What changed (motivation → approach → change)

Root cause: the model wrote prose about facts and checks tried to catch it afterwards. The fix moves every fact to gateway code and every judgement to a ledger record the conductor writes.

  • Welcome (dashboard/warm_task_brief_welcome.py, dashboard/warm_task_brief.py). Four parts in order, in the owner's language: Goal (the ledger goal text); Needs you (one plain line when empty; otherwise, per item the Board's own awaits_person test says waits, its state, its PR state when read live and "checked HH:MM", then the why, every option with what it leads to, and the pick with its reason; with no record, a plain line saying the reason and options were not written yet; question and change cards and a reply's open choices too); Being done now (one line per started item: ledger status, PR state only when read live, "checked HH:MM"); Comes after (items not started, labelled plan). No model call and no wake turn. The row is an ordinary Bot message (no meta.brief) with main's copied [OPTIONS:] line. A quiet return stays one line. After the reads, the ledger and the waiting list are read again before the post; a change renders again (at most three times).
  • Ask record (work_ledger_record action=ask). New conductor action with item_id, what, why, options: [{label, outcome}], pick, pick_why; the store stamps written_at, refuses a pick outside the labels, and decide/close clear it. Its event kind is ask, not decision, so it never reads as an answer. It rides work/recorded (entry field, baseline, fold), so a rebuild restores it. The goal-conductor skill and the shared conductor prompt say to record it in the same turn a question is relayed to the person.
  • Board updates when the welcome is generated. Each good live PR read is recorded once as object/observed (producer warm_brief) in the thread's own unit, no extra GitHub call. A new objects slot fold keeps the newest read per subject and producer (64 subjects, 2 KiB facts, out-of-order reads never overwrite newer ones). /api/crew-board (owner-gated, unchanged) adds pr_facts; the Crew Board page draws its tags with "read HH:MM".
  • Preparing hint (components/chat/PreparingReplyBubble.tsx, lazy-loaded). Raised before the first GitHub read, inside the response bubble in both the Ultra Mode chat (ChatTranscriptView) and the Bot pane (ChatPane), replaced in place by the welcome.
  • Ledger text is inert. Every ledger string is redacted whole in every scan view before any cut, collapsed, backticks neutralised, control markers defanged, links and addresses made inert, HTML escaped; value words are case-folded; times are strict HH:MM. No user-facing text says "conductor" or "指挥" (docs/decisions 2026-10-30).
  • Removed: the one-call model brief, its prompt, grounding scans and the wake module (warm_task_brief_wake.py), with their tests.
  • RFC and specs. RFC section 3.11 and a dated amendment bullet; session, crew-mode, crew-log core/emitter/session-types, ledger reference and feature-map rows updated.

RFC: rfc-delayed-task-brief.md is partial on main; section 3.11 is added by this PR at the owner's direction.

Backwards compatibility

Compatible: a stored brief keeps drawing as its card through WarmBriefCard; an item with no ask is stored and read exactly as before (the field is written only when set); object/observed widens a closed producer vocabulary and adds no required field; the new ask action and event kind widen closed vocabularies.

Owner ruling (2026-11-12): the ask write is its own deferred tool on kirocrew-work, work_ask_record (item_id, what, why, options[{label, outcome}], pick, pick_why), sending the record route's existing ask write. It is not in JOB_TOOLS, so it costs no schema tokens until a conductor loads it. work_ledger_record keeps main's description and schema, and the test/test_job_tools.py budgets (7,600 / 7,800) are unchanged. The goal-conductor skill and the conductor prompt say to load and call it in the turn a question is relayed to the person; the goal conductor's grants include it. The welcome and the Board read the same stored record.

Reader: src/kiro_crew/crew_log/projection_work.py:_work_apply -- crew page -- applies the new ask field and clears it on decide/close (test_work_ledger_projection.py)
Reader: src/kiro_crew/work_ledger_records.py:WorkItem.from_dict -- crew page -- reads ask through coerce_ask; absent reads as None (test_work_ledger.py)
Reader: src/kiro_crew/dashboard/warm_task_brief.py:_run_warm_task_brief -- chat -- the only caller of the removed warm_task_brief_wake module and of the model brief; it now writes the welcome itself (test_warm_task_brief_welcome.py)
Reader: website/src/components/chat/WarmBriefCard.tsx -- chat -- unchanged; new welcome rows carry no meta.brief and draw as an ordinary message (ChatPage.views.test.tsx)

Tests

  • test/test_warm_task_brief_welcome.py (new): the four parts and their order; the record with every option and the pick; the plain not-written line; a pick outside the labels hidden; PR named only when read, failed read "not read since"; case-folded value words, bad first; a reply's choices not listed twice; control markers, links, images, backticks and a cap inside an address all inert; a key redacted whole before the cut; every language's words and no "conductor"; a card answered during the reads re-renders; a remote worker's binding kept; the objects fold and warm_brief producer; pr_facts.
  • test/test_work_ledger.py, test/test_work_ledger_projection.py: ask stored, refused when malformed, cleared by decide and close, carried through a rebuild.
  • test/test_warm_task_brief*.py: adapted to the gateway welcome (render seam test/warm_brief_render_seam.py); model-path tests removed with the model path.
  • Website: prFactTags.test.ts, CrewBoardPage.focus.test.tsx, ChatPane.preparing.test.tsx, ChatPage.views.test.tsx.
  • Local: touched Python tests (635 passed on the rebased head), touched vitest files (25 passed), tsc, eslint, black, mypy. Local GPT 6.1 and Opus reviews ran once; their blocking findings (a cap inside an address, a ledger backtick, the remote worker binding, no re-read before post) are fixed in this commit.

Proof: not provable -- other: feature change, not a fix; the tests pin new behaviour rather than a reverted defect

Manual verification

Captured against the capture harness (members page) at 1280 and 360 px with every frame asserted. The two "after" texts were rendered by the gateway's own welcome.render and _clean. Not run against a live gateway.

Screenshots / video

Before (main): the model-written brief card

before, 1280

Preparing hint inside the reply bubble

preparing, 1280

After: one item needs you (why, options with outcomes, pick; chip row unchanged)

after with needs-you, 1280

After: nothing needs you

after with nothing waiting, 1280

360 px

before, 360
preparing, 360
after with needs-you, 360
after with nothing waiting, 360

After: a welcome that needs the owner, with the reply line just above the chips ("点下面的选项,就是发出你的回复。", committed under temp-screenshots/warm-welcome/)

needs you with the reply line, 1280
needs you with the reply line, 360

Crew Board: an item row with the welcome's PR fact tags (one bad tag "has conflicts" in the warn tone, one good tag "checks 12 passed, 0 failed", and a read from an earlier day shown with its date)

Crew Board PR fact tags, light, 1280
Crew Board PR fact tags, dark, 1280
Crew Board PR fact tags, dark, 360

When the welcome carries the copied [OPTIONS:] row, one line just above it says a chip only sends a reply: "Tap an option below to reply." (zh "点下面的选项,就是发出你的回复。"), in all seven welcome languages.

Related Issues

Fixes #32844 (the preparing hint now sits in the reply bubble). Refs #25306. Replaces #33032 (closed).

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

Contribution License Agreement

Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.

@iamwhatever
iamwhatever requested review from a team and KiroCrewCaptain as code owners November 9, 2026 08:32
@iamwhatever
iamwhatever requested a review from Zedmor November 9, 2026 08:32
@github-actions

github-actions Bot commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Sensitive Change Review

Sensitive files changed:

  • src/kiro_crew/security_posture.py

  • @iamwhatever posted a complete reasoning template for 7a50589.

@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 9, 2026
@github-actions

github-actions Bot commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ human override accepted

@iamwhatever overrode this lane for 7a5058908b91dd04838ee827cc96697cb85538b3 via /ai-review override (this commit only).

@github-actions

github-actions Bot commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 7a5058908b91dd04838ee827cc96697cb85538b3 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks. Moving every fact into gateway code and every judgement into a ledger record removes the false model sentences the audit found. Rendered evidence is attached, and no reader outside the patch uses the removed model-brief path.

Notes

  • The Board pr_facts fold and the preparing hint are separate surfaces bundled into one PR. The owner asked for one PR, but each adds its own way to break, so splitting them out would make a revert cleaner.
  • The welcome's "Needs you" detail only works if the conductor writes the ask record. Telling it to in the prompt doesn't make it happen. Track how often the "not written yet" line shows up, so you can see whether the record is actually being written.

[DESIGN-REVIEWED] 7a50589

Description read

[DESCRIPTION-READ] 12230faf81c26516dbfac3975360340e45220ae393cb206c4f10e4245d4542c3

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 .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS

UX-level review of 7a5058908b91dd04838ee827cc96697cb85538b3 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

UX-Verdict: PASS

Nothing blocks. Screenshots show every changed surface, and the blind reader read the welcome, the preparing bubble and the Board PR tags correctly.

Notes

  • When the welcome reads a PR live, each line shows the same time twice: "GitHub 06:52,06:52 核实" (shot-09). The blind reader said: "Seeing the same time twice in a row looks like a mistake." When the github {at} time equals checked_at, drop the per-item checked. Or print "checked HH:MM" once under the recap.
  • An item with ledger status done shows "报了完成" / "reported done" under 正在做 / Being done now (shot-09). The reader found that odd. List reported-done items under their own heading, or reword the status to "reported done, not closed yet".
  • The Board row shows "PR fix(sidebar): let a person reorder a nested subfolder, not only an agent #12208" next to kirodotdev/KiroCrew#33701 with its tags (shot-11). The reader "can't tell which number is the PR". Prefix the facts label to say this is the PR the tags describe, e.g. "read for fix: name the package to install when an installer tool is missing #33701".

[UX-REVIEWED] 7a50589

Description read

[DESCRIPTION-READ] 43ac794bda68ba34ad236f8a2021038de087dee4a1d96db714fd168d3e002dfe

That sha256 names everything this verdict read: the PR description, plus the 16 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 .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 7a5058908b91dd04838ee827cc96697cb85538b3 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 7a50589

Verdict parsed from the review's SHA-scoped output markers for commit 7a5058908b91dd04838ee827cc96697cb85538b3.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 7a5058908b91dd04838ee827cc96697cb85538b3 false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings (all downgraded on adjudication)

GPT 6.1 flagged blocking issues on 7a5058908b91dd04838ee827cc96697cb85538b3; Opus 5.5 adjudication downgraded every one of them. Downgraded: the remedy each one needs is disproportionate to the harm it prevents. Adjudication downgraded all 1 blocking finding(s): the remedy each one requires is disproportionate to the harm it prevents.

This comment is updated in place on each push.

Review details

BLOCKING -- src/kiro_crew/dashboard/warm_task_brief.py:1102 -- Final async composition lets dismissed cards reappear in the welcome

await _compose(state, slot, settled, facts.timeless(live), language, _DIGEST_EPOCH)
Trigger: The owner dismisses a failed change card through api_cards_dismiss while _run_warm_task_brief awaits this final composition’s threaded reads. Dismissal removes the card from the waiting list without cancelling the welcome or starting a turn.
Outcome: Publication uses the previously rendered content and captured settled.waiting, so “Needs you” still lists the card the owner dismissed. The final slot-identity/running check does not detect this change.
Anchor: defect
Fix: Finish asynchronous reads first, then refresh the waiting list and synchronously render content and digest before posting without another suspension.
Sites: src/kiro_crew/dashboard/warm_task_brief.py:1093 also suspends after the last snapshot; src/kiro_crew/dashboard/warm_task_brief.py:1195 renders the captured waiting list.

[BLOCK-MERGE-DOWNGRADED] 7a50589
[GPT-REVIEWED] 7a50589

Adjudication (Opus 5.5) — is blocking on each finding proportionate?

F1 — LOW. For this to happen, the owner has to dismiss the card in the few milliseconds the gateway spends in threaded memory reads (asyncio.to_thread(reads) at warm_task_brief.py:1186, reached from :1093 and :1102), after the last _settled re-read (:1090). The final check at :1105 only looks at slot.running and _same_slot (:720), so it does not catch a dismissal. When it happens, the welcome lists a card the owner just dismissed. The dismissal itself still takes effect (change_cards.py:1063); only this one welcome message is stale, and the owner can see that, since they just made the dismissal. Recovery: the next welcome reads the waiting list again from scratch (_settled at :1071/:1090), so the stale line does not come back. Nothing is corrupted or lost.
Fixing it means restructuring _compose/_render so the reads finish before rendering, re-reading waiting and rendering synchronously. That adds a new ordering contract to this one function, and every future reader of it has to keep that contract in mind. That cost is larger than a rare, visible, self-correcting stale line in one message.

[ADJUDICATION] 7a5058908b91dd04838ee827cc96697cb85538b3 total=1 uphold=0 downgrade=1
DOWNGRADE F1 src/kiro_crew/dashboard/warm_task_brief.py:1102 reason=disproportionate-remedy
[GPT-ADJUDICATED] 7a5058908b91dd04838ee827cc96697cb85538b3

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 7a5058908b91dd04838ee827cc96697cb85538b3 false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 9, 2026
@iamwhatever
iamwhatever force-pushed the feat/warm-welcome-four-part branch from 8f97a35 to 39096ac Compare November 9, 2026 09:02
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Nov 9, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 39096ac
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 9, 2026
@iamwhatever
iamwhatever force-pushed the feat/warm-welcome-four-part branch from 39096ac to eb06e66 Compare November 9, 2026 09:52
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: eb06e66
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Nov 9, 2026
@iamwhatever
iamwhatever force-pushed the feat/warm-welcome-four-part branch from eb06e66 to 54b3850 Compare November 9, 2026 10:25
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Nov 9, 2026
@iamwhatever
iamwhatever force-pushed the feat/warm-welcome-four-part branch from 54b3850 to 86eb770 Compare November 9, 2026 10:37
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 86eb770
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 9, 2026
@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Nov 12, 2026
@iamwhatever
iamwhatever force-pushed the feat/warm-welcome-four-part branch from 8fb9e51 to c4c47a4 Compare November 12, 2026 17:56
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: c4c47a4
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Nov 12, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 1ac335f
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 49e1639
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: e6d911e
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles e6d911e owner-decision: Owner accepts the product-shape finding; RFC change ships in this PR by owner direction. Decision: #33805 (comment)

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@iamwhatever recorded a owner-decision release (class owner-decision) for the first-principles AI finding on e6d911e2f525b95f6b6727f32c2a1b0853f5050c.

Owner accepts the product-shape finding; RFC change ships in this PR by owner direction. Decision: #33805 (comment)

This decision applies only to this commit. A new push requires a new judgment.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: bdd6ddc
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: a5a4d53
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 7b1391c
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

UX Review on a5a4d53d31c1: the blocker and the evidence gap are fixed in 7b1391c1da508e8fa644c304f6a29baaef62cf89. Dispositions for the three non-blocking notes, with no code change:

  • "Preparing your catch-up…" is a guess for some readers. Deferred. The string is the hint main already ships for the preparing state (section 3.9.7), and this PR only moves it inside the reply bubble. Rewording it is a copy change for every surface that shows it, outside this PR's goal.
  • Name the PR before its facts on each line. Deferred. Each item line already reads title (state, repo#N: facts, GitHub HH:MM, checked HH:MM), so the PR is named before its facts. Splitting two PRs that one reason mentions needs the ask record to name its PRs, which it does not carry.
  • Show "Decided: …" once an ask is answered. Deferred. The welcome shows only what waits now, by the owner's ruling of 2026-11-09 ("facts only"). An answered ask leaves "Needs you", and the decision stays in the thread and on the Board.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 1121cfb
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: d0e878c
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: dbc1531
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: f01bae3
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 35202b4
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

…gateway

The warm welcome is now a short, plain recap the gateway writes in the
owner's language, with no cards, pictures, widgets or model call: the
goal; what needs the owner, each option explained with what it leads to
and the pick with its reason; what is being done now with its state and
check time; and what comes after. The judgement comes from a new
conductor action, work_ledger_record action=ask, which decide and close
clear. Each live pull-request read is recorded as object/observed
(producer warm_brief) into a new objects slot fold that the Board draws.
The preparing hint sits inside the reply bubble in both chat surfaces.

RFC rfc-delayed-task-brief.md section 3.11 records the owner's ruling of
2026-11-09. Replaces closed #33032.
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 7a50589
  • What rule changed: src/kiro_crew/security_posture.py adds one entry, dashboard/warm_task_brief_check.py, to NON_EGRESS_REDACTION_MODULES. No sandbox rule, command floor, scrubber or redactor changed.
  • Before -> after (what was allowed/blocked, now): before, warm_task_brief_check.py called no redactor. Now it holds redact_whole, a pure helper that returns a redacted copy and emits nothing. Its only caller is dashboard/warm_task_brief.py, which is already a registered redaction sink and posts the welcome. Without the entry, the posture test would refuse a redactor call site that is neither a sink nor allowlisted.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): the security panel undercounts one redaction call site. Nothing is allowed or blocked differently; the welcome still redacts every ledger string whole before any cut and scans the exact posted text.
  • Evidence checked (which sandbox result proves it): no sandbox behaviour changed, so no sandbox run applies. test/test_security_posture.py (omission detection: every redactor call site is a registered sink or allowlisted) passed locally on this head's code.
  • No regression:
    • Tools still work (which tools/commands were run, result): N/A: no tool, command or sandbox path is touched; the change is one allowlist string.
    • Backward compatible (old config / old data / old callers still work, how checked): yes; the set only grows, and no config or stored data reads it.
    • Existing tests (which suites passed on this head): test/test_security_posture.py, test/test_warm_task_brief*.py, test/test_work_ledger*.py passed locally.
  • How to undo: revert this PR; the entry goes with the helper it describes.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles 7a50589 owner-decision: Owner accepts the product-shape finding; RFC change ships in this PR by owner direction. Decision: #33805 (comment)

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@iamwhatever recorded a owner-decision release (class owner-decision) for the first-principles AI finding on 7a5058908b91dd04838ee827cc96697cb85538b3.

Owner accepts the product-shape finding; RFC change ships in this PR by owner direction. Decision: #33805 (comment)

This decision applies only to this commit. A new push requires a new judgment.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Warm welcome: no indicator while the brief is prepared, so the owner types and cancels it

4 participants