Repository navigation
fix(release): backport two dashboard and Bot steering fixes to 0.12.0 line - #31422
Conversation
… every Bot turn (#31131) * fix(members): load steering up to the document cap instead of failing every Bot turn A Bot inherits global steering by default. With more than 64 steering files, the essential-context reader raised 'too many documents' and every Bot turn failed, while main chats kept working. Steering past the cap is now left out unread and named in an in-band note and a warning, the same way oversized and unreadable steering already is. The first files in sorted order are kept. A member's own sources (persona, SOUL.md, declared resources) take the last kept steering file's slot before they would be refused, so they are never the ones dropped. The cap itself is unchanged. The native launch list and the projected KAS resource snapshot get the same treatment so they do not refuse the session start either. * fix(members): never evict a core source to make room at the document cap A steering file reused as the Bot's file:// persona is marked core, but it stayed in the evictable steering list, so a later source (the template's context settings) could pop it and the persona was lost. Marking a source core now removes it from that list, and an evicted path leaves 'seen' so a later source naming it loads it again. Documents the keep-first-64 rule and the over-cap note in the memory spec. (cherry picked from commit 73af860) Conflict resolution (release/0.12.0): src/kiro_crew/member_essential_context.py, one hunk in _resource_paths. Main's side also changed the `escaping` line (`else left_out` instead of `else None`), which comes from a later main PR this branch does not carry. The release `escaping` line is kept, and only this pick's change is applied: the `steering_glob` local and its use in `left_out`. Every other hunk applies unchanged.
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing to check. The PR adds no control, label or state: it keeps unsent draft text across a reload, and the steering cap is a backend fix. [UX-REVIEWED] 79349e1 Description read[DESCRIPTION-READ] c251b5d18bfb8f6ec51d433ce428bdc82b1792c8e86e533b9d6dc6184424cd40 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 |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: this backports two fixes already merged on [DESIGN-REVIEWED] 79349e1 Description read[DESCRIPTION-READ] 45a0077ed4c1a33afd75bc28eff6de401ae8b52f936e24588c4fb328f2fada75 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. Both picks are fixes for linked issues (#31296, #31128) that already merged on What this change shipsInventory (6 items) — 6 justifiedIntent: ship two
[FIRST-PRINCIPLES-REVIEWED] 79349e1 Description read[DESCRIPTION-READ] 36f28ec805756dcd1401898575cd2ec2403727be416d45183c65ecc4c9022d7e That sha256 names the PR description this verdict read. 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: |
Problem / Motivation
Goal: Bring two dashboard and Bot fixes that merged on
mainafterv0.12.0ontorelease/0.12.0, one commit per fix, so the 0.12.1 patch ships them and each one can be reverted alone.Why it matters
Not a goal
main; nothing is changed only here.CHANGELOG.mdchange. Cuttingv0.12.1-rc.Nand the bare tag follow docs/build/release.md and stay with the maintainers.What changed (motivation → approach → change)
Each main squash commit was cherry-picked with
-xontorelease/0.12.0@8da4c55e6e, in this order, one commit each.090dad8cf13484537c0173af86099f79349e1b8dThe only hand resolution was one hunk in
_resource_pathsinsrc/kiro_crew/member_essential_context.py:steering_globleft_outsteering_globsteering_globescapingelse Noneelse left_outelse NoneThe
escapingline stays as release has it becauseelse left_outbelongs to #30722, not #31131 (git log -S'else left_out'names only #30722). Taking it would carry part of an unpicked fix. Every other #31131 hunk applies unchanged, including theover_cap=... if steering_globargument that uses the new name. The commit message records the resolution.Backwards compatibility
Compatible: each change is the one already merged on
main, and each PR's own compatibility note applies unchanged.Removes nothing: no pick hides, deletes or renames anything 0.12.0 accepts.
Tests
Proof: not provable -- other: a backport of main commits whose proofs ran on their own PRs
release/*, so the related set was run here: the addedtest_member_essential_context_steering_doc_cap.py(7 passed) plus every test file that importsmember_essential_context, 21 files, one pytest process,-n 0. 1042 passed. 2 tests fail, and both fail the same way on cleanrelease/0.12.0@8da4c55e6e, so neither comes from this diff: two intest_folder_steering_provider_dedup.py.useChatPaneDraftfrontend tests, including the added reload test: 4 passed.Manual verification
N/A - every pick was verified on its own main PR. The 0.12.1 RC soak is the end-to-end check.
Screenshots / video
Why no screenshot: #31308 only saves the composer text when the page is about to be hidden, so nothing renders differently, as its own PR states.
Related Issues
Backport of #31308 and #31131 to the 0.12 line for 0.12.1. Source issues: #31296, #31128.
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 two fixes already on
main. Each source PR has its own harvest.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.