Repository navigation
chore(release): start the 0.9.0 insider line - #18356
Conversation
…GELOG) Declare the release/0.9.0 branch identity as the dual-valid RC spelling the version numbering policy requires on an insider release branch: all three version files move from the bare 0.9.0 main carries to 0.9.0-rc.1. Add the ## [0.9.0] - 2026-10-09 section, written from the range 35fac20..69dc0ce minus the commits backported to release/0.8.0. Every earlier section is unchanged.
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Check before tagging: the post-merge step says to push What this change shipsInventory (2 items) — 2 justifiedIntent: open the 0.9.0 insider line so its builds identify as release candidate 1 and its release notes exist at the cut (an ADDITION; this is release chore work).
[FIRST-PRINCIPLES-REVIEWED] 2aaf246 Description read[DESCRIPTION-READ] 90463be5b185645633b48188e7f852fc77f5eed01707f200ff4711b390c778d0 That sha256 names the PR description this verdict read. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Check that the tag scheme is settled: release.md's policy says to tag [DESIGN-REVIEWED] 2aaf246 Description read[DESCRIPTION-READ] 1372c5fc5d2df15c6ceb597834e09d044ca35e997eb46e5daf0b0e3381f1c631 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
GPT 6.1 Review — ✅ 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: |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Check that the About dialog and the update prompt show [UX-REVIEWED] 2aaf246 Description read[DESCRIPTION-READ] bb02512abd58a53800461b6919c45870b913a8e82aa19891b945c8e2b5bfca28 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 |
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: |
What
Start the 0.9.0 insider line on
release/0.9.0:0.9.0thatmaincarries to the dual-valid RC spelling0.9.0-rc.1:src/kiro_crew/__init__.py,pyproject.tomlandwebsite/electron/package.json. Thewebsite/electron/package-lock.jsonroot still reads0.7.0, as it does onmainandrelease/0.8.0; it is not one of the three declaration files and is left alone.CHANGELOG.mdgains## [0.9.0] - 2026-10-09above[0.8.0]. Nothing else in the file changes.Why
The version numbering policy in
docs/build/release.mdsays an insider release branch declaresX.Y.Z-rc.Nin all three files, and the RC runbook says the branch carries its## [X.Y.Z] - <date>section, with no[Unreleased], at the RC cut.release/0.9.0was cut frommainand still declared the bare0.9.0with no 0.9.0 section.CHANGELOG range
35fac2012914f17dad3195a1f472d6df1c72a6ee(merge-base ofrelease/0.8.0and the branch head)..69dc0ceee3d2812c43078cebbd2ae8431ee245a4, first-parent: 1636 commits.release/0.8.0subjects and bodies) and are left out.test, 106docs(including the notes-only CHANGELOG port chore(version): port the final 0.8.0 CHANGELOG section to main #18282), 87refactor, 28ci, 9chore, 1build, 1style.feat, the restfix,perfand untyped). Reverted pairs are left out: the crew mode toggle (feat: add crew mode toggle to chat sessions #16088, reverted in revert: back out crew mode toggle (#16088) #16257), refusing to close a session with a running worker (feat(sessions): confirm before closing a tree whose workers are still running #17317, reverted in revert: refuse close when any worker under a session is still running (#17317) #18003), and the transcript scrolling under the composer glass (fix(chat): keep floating composer chrome off the transcript and hero #15820, reverted in revert: transcript scrolls under the composer glass again (#15820) #16291).docs/feature-map/README.mdandsrc/kiro_crew/docs/settings-registry.generated.jsonat the branch head, and against commit bodies where those were silent. The snapshot read fix fix(dashboard): skip the context snapshot write when its read failed before restore #18277 is in Notable fixes.The section follows
docs/build/changelog.md: an opening paragraph, twelve subsections of at most three one-sentence bullets, two marked(Preview), and a Notable fixes section of three short paragraphs. It has noBefore you upgradeand noContributorssection, and no em or en dash.The 0.8.0 section
Unchanged. The
[0.8.0]section on this branch is already the final one: extracted from## [0.8.0]to the next## [heading, it is byte-identical to the one onrelease/0.8.0at4d205dadc142500dd3f0e89de28ce70304b72d7c(529 lines, empty diff).Checks run on the head of this branch:
CHANGELOG_BASE_REF=origin/release/0.9.0 python3 scripts/check_changelog_history.py:12 shipped section(s) at origin/release/0.9.0 are intact at head OK.python3 scripts/check_changelog_history.py:no draft sections OK.pytestontest_stable_version_display.py,test_stable_release_gate.py,test_changelog_parse.py,test_changelog_history_gate.py,test_release_channel.py,test_build_version_override.py,test_cli_version_fastpath.pyandtest_kiro_crew.py: 276 passed.After merge
A maintainer pushes an annotated tag
v0.9.0-insider.1on the merge commit, with the messageKiroCrew 0.9.0 insider candidate 1. Do not pushv0.9.0-rc.Ntags on this base: anrc.Nand aninsider.Ntag with the same number map to the same wheel version.Pattern harvest
Not generalizable: release branch cut with no defect behind it.