Skip to content

fix: backport 0.12.1 fast batch v2b-4 (9 user-facing fixes) - #33019

Merged
buluoray merged 9 commits into
release/0.12.0from
crew/slack/0.12.1-fast-v2b-4
Nov 8, 2026
Merged

buluoray merged 9 commits into
release/0.12.0from
crew/slack/0.12.1-fast-v2b-4

Conversation

@buluoray

@buluoray buluoray commented Nov 8, 2026

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Bring nine fixes for problems a 0.12.0 user can hit, merged on main after v0.12.0, onto release/0.12.0, one commit each, in a PR of its own.

Why it matters

Each fix changes behaviour a 0.12.0 user can run into today. Every pick applied to release/0.12.0 cleanly, so the code each one replaces is on release.

PR Area What a 0.12.0 user sees today
#31004 session idle-session cleanup runs from boot on every gateway
#30986 security tool-call pause redaction uses an exact word match
#31304 Slack Removing a Slack allowlist user_id entry actually removes it
#30482 UI Holding Enter no longer confirms webhook revoke or delete
#32454 deploy A malformed profiles.json makes deploy routes return 500
#32478 bots/work-ledger Work-ledger rebuild always refuses with crew_log_incomplete
#32489 voice Mic turns on for 15s even with speech-to-text set Off
#31779 apps Deeply nested registry cache file crashes apps with RecursionError
#32238 artifacts Opening an SVG artifact restyles the whole dashboard

Not a goal

  • No fix that is not already merged on main; nothing is changed only here.
  • No version bump, no tag, no CHANGELOG.md change.
  • Rows whose pick conflicts; they go in a later batch.

What changed (motivation → approach → change)

Each main squash commit was cherry-picked with -x onto release/0.12.0 @ 2fc09a0f42, in merge order on main, one commit each. Every pick applied cleanly, and every commit keeps its original author.

# PR main commit release commit
1 #31004 start the cleanup loop at boot on every gateway 0846b6a969 56cf197a57
2 #30986 use an exact check for the word shown at a tool-call pause a1ff905a97 9eba07fd60
3 #31304 remove a user_id allowlist entry from the slack config d449b7e44f 844dabc6a8
4 #30482 ignore a held key on armed Revoke and delete, and return focus after Keep f05f9fb1cf 8095d18b83
5 #32454 degrade a profile registry json.loads refuses outright 32ce0d8c09 c6dd313c96
6 #32478 stop counting the checkpoint folder as a broken crew-log unit 535df79ba1 59966d258f
7 #32489 do not warm the mic when speech-to-text cannot run 87266ea666 a105a286cb
8 #31779 treat a deeply nested registry cache file as absent 94de517a9b cedaa6b194
9 #32238 keep an svg artifact's own styles from restyling the dashboard f26fb971d0 137ff18b86

Picks whose co-author trailers carried internal addresses have each trailer rewritten as a plain "Co-author on main: " line, because the commit guard refuses the address.

#30664 and #30861 are left out: review on another release batch found defects in them that also apply to main, so they wait for fixes there.

Backwards compatibility

Compatible: each change is the one already merged on main, and its compatibility note applies unchanged.

Removes nothing: no setting, key or caller is removed.

Tests

Proof: not provable -- other: a backport of main commits whose proofs ran on their own PRs

  • CI does not run the backend suite on PRs into release/*, so the direct test modules were run here: the Python test files these picks touch gave 828 passed.
  • Frontend: tsc -p tsconfig.app.json --noEmit passes; the 3 vitest files these picks touch gave 38 passed; lint:i18n passes at the same count as clean release.
  • flake8, isort, the black gate and mypy src/kiro_crew pass. i18n:check without a base reports the same findings class as a clean release/0.12.0 on this host, and fewer of them.

Manual verification

N/A - each fix was verified on its own main PR. The 0.12.1 RC soak is the end-to-end check.

Screenshots / video

Why no screenshot: the frontend picks carry the same rendered change as their main PRs, which show it.

Related Issues

Backport to the 0.12 line for 0.12.1 of #31004, #30986, #31304, #30482, #32454, #32478, #32489, #31779, #32238.

Sibling sweep

Siblings: none -- each source PR carried its own sibling sweep; this PR only carries those fixes.

Pattern harvest

Not generalizable: this PR only carries fixes already on main. The source PRs have their own harvest.

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 and others added 9 commits November 8, 2026 07:38
Co-author on main: Zejiang Guo
(cherry picked from commit 0846b6a)
A reply's last word stays on screen while a tool runs only when it is
one plain word ending a sentence. Decide that with an exact check of
each character, and pin, for every shape the redactor removes, that a
value written right after such a word, with the pause there or at any
dot inside the value, never reaches a row, the stream or the saved
reply unredacted.

(cherry picked from commit a1ff905)
The loader reads a slack.allowed_users entry such as {"user_id": "U1"}
as user U1, but _update_config_list matched only on slack_id. So the
Deny and Remove buttons dropped U1 from the running allowlist and
reported success while config.json kept the entry, and U1 was allowed
again after a restart. The matcher now also matches a user_id key
standing in for an unusable slack_id, for remove and for the
duplicate check on add.

(cherry picked from commit d449b7e)
…32454)

load_registry promises that one bad profiles.json cannot 500 every route
that reads the registry. Its three reads caught json.JSONDecodeError and
UnicodeDecodeError only, so a file nested past the interpreter's
recursion limit (RecursionError) or holding a number past the integer
digit limit (a plain ValueError) escaped every reader. The locked writer
caught ValueError but not RecursionError, so a nested file raised out of
a profile save instead of being moved aside.

Each read now catches ValueError and RecursionError, and the locked
writer sets a nested file aside to <name>.corrupt-<utc> before the save,
as it already does for truncated JSON.

Builds on #29252.

Closes #29089

(cherry picked from commit 32ce0d8)
…32478)

The work-ledger rebuild refuses while any session crew-log unit has an
unreadable header. The count walked every child of the session root,
including the projections/ folder the session tree saves its checkpoint
in, so every rebuild on a home with that checkpoint refused with
crew_log_incomplete. Skip that folder by name, as the session tree's
seed scan already does; a real unit with an unreadable header still
refuses.

Closes #30321

(cherry picked from commit 535df79)
Three readers of the agent-writable registry cache directory (the
manifest cache, the external registry cache and the clone-failure
record) let RecursionError from json.loads escape. A body nested deeply
enough to exhaust the decoder's stack then dropped a registry from the
App Store listing or ended the store-art prewarm batch, instead of
reading as "no record" as every other malformed body does. Each
reader's except now lists RecursionError, matching the readers #22300
already guarded.

Closes #22359

(cherry picked from commit 94de517)
…32238)

An SVG artifact can carry its own stylesheet, and when it was drawn in
the viewer or on its library card that stylesheet could change the rest
of the dashboard and its printouts. Those stylesheets are now left out,
so an SVG styled that way shows without its stylesheet instead.

Closes #31600

(cherry picked from commit f26fb97)
@buluoray
buluoray requested review from a team and KiroCrewCaptain as code owners November 8, 2026 07:50
@buluoray
buluoray requested review from Zedmor and removed request for a team November 8, 2026 07:50
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 8, 2026
@github-actions

github-actions Bot commented Nov 8, 2026

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS

UX-level review of 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

UX-Verdict: PASS

Nothing blocks: the diff adds or changes no user-visible control or label. The changes are key-repeat guards, focus return after Keep, skipping the mic warm-up, and stripping <style> from inline SVGs.

Notes

  • FORBID_TAGS: ['style'] in SvgViewer and ContentThumb changes how SVG artifacts look. An agent-made SVG that colours its shapes through <style> classes will now render with default black fills and no styling. It would help to inline those rules as attributes (or scope them under the SVG) instead of dropping them, so artifacts keep their look.

[UX-REVIEWED] 137ff18

Description read

[DESCRIPTION-READ] c3d844c11f7d92c2be3489315b2dba7375429edb499463eb90e7883910c5b337

That sha256 names everything this verdict read: the PR description, plus the 3 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 8, 2026

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

This is a clean backport of nine fixes already merged on main. No pick adds anything beyond what its source PR merged.

Notes

  • Item 1 — undeclared: the diff deletes test_a_non_blocking_start_leaves_the_check_to_a_background_task ("A non-blocking start returns before its housekeeping") and _start_cleanup_for_waiting_deletes, and the description never mentions either. fix(session): start the cleanup loop at boot on every gateway #31004 on main is the decision that reversed that pin, so the deletion is backed, but a reader of this PR does not learn about it.

What this change ships

Inventory (9 items) — 9 justified

Intent: give 0.12.x users nine user-facing fixes already on main, via 0.12.1. FIX.

  1. The idle-session cleanup loop starts at boot on every gateway, not only when deletes are waiting — justified
  2. Pausing at a tool call checks the shown word exactly instead of with a regex — justified
  3. Removing a Slack allowlist user_id entry actually removes it — justified
  4. Holding Enter no longer confirms webhook revoke or delete; focus returns after Keep — justified
  5. A malformed profiles.json degrades instead of making deploy routes return 500 — justified
  6. The checkpoint folder no longer makes a work-ledger rebuild refuse with crew_log_incomplete — justified
  7. The mic does not warm up when speech-to-text is Off or unavailable — justified
  8. A deeply nested registry cache file is treated as absent instead of crashing with RecursionError — justified
  9. An SVG artifact's own styles no longer restyle the dashboard — justified

[FIRST-PRINCIPLES-REVIEWED] 137ff18

Description read

[DESCRIPTION-READ] f662ac0027d477d4696cc8a34c941e7a14e9a6757a62674e76daed78ffa6c556

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

github-actions Bot commented Nov 8, 2026

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Clean-pick backport of nine fixes already on main; no new design, surface, or broken reader.

Notes

  • test_kas_released_sessions.py reverses the old assert_not_called pins on starting the cleanup loop at boot. That reversal is the decision fix(session): start the cleanup loop at boot on every gateway #31004 already made on main, so it isn't treated as a gap closed here.
  • This PR puts the fix for "idle-session cleanup now runs from boot on every gateway" on the release line. That changes boot behaviour on every 0.12.x gateway. The RC soak named under Manual verification is the only end-to-end check of it, so name that soak explicitly in the 0.12.1 release checklist.

[DESIGN-REVIEWED] 137ff18

Description read

[DESCRIPTION-READ] e8912c8d98d81f0e38448dc465b8792fe5b4fc209daf182dee60c8ce50aa0507

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 8, 2026

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 137ff18

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 8, 2026

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 137ff18

Verdict parsed from the review's SHA-scoped output markers for commit 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 137ff18b8691cfd2d345b49cbfbe83bcc7dbaf99 false-positive: <reason, 20 to 500 characters>

@buluoray
buluoray merged commit 3d0dc11 into release/0.12.0 Nov 8, 2026
20 of 23 checks passed
@buluoray
buluoray deleted the crew/slack/0.12.1-fast-v2b-4 branch November 8, 2026 08:00
@github-actions github-actions Bot removed the readiness: checking Automated validation is still running label Nov 8, 2026
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.

4 participants