Repository navigation
refactor: 0.13 backport, rename Ultra Mode to conductor mode (#36924, #37095) - #37506
Conversation
Sensitive Change ReviewSensitive files changed:
How to clear this check:
Only the author's comment counts this way. A reviewer can instead |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS This backports a rename to 0.13 that maintainers already ruled on (#36869, #37086). Behaviour does not change, and the removed old-spelling compatibility only ever applied behind a preview that is off by default. What this change shipsInventory (9 items) — 9 justifiedIntent: 0.13 users meet the preview feature under its final name, conductor mode, and never see the name change again in 0.14. FIX (a backported relabel).
[FIRST-PRINCIPLES-REVIEWED] cfa522f Description read[DESCRIPTION-READ] f9a3a88d1eee95353f954c1c60016f5b62129dd37e839c9bbc3f157a19faa8e5 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: |
Human judgment recorded@buluoray recorded a not-applicable release (class
This decision applies only to this commit. A new push requires a new judgment. |
Human judgment recorded@buluoray recorded a not-applicable release (class
This decision applies only to this commit. A new push requires a new judgment. |
UX Review (Opus 5.5) — ✅ human override accepted@buluoray overrode this lane for |
The crew feature's final product name is conductor mode. Every user-facing place that said Ultra Mode says conductor mode: UI strings in all locales (kept untranslated), settings labels and their generated search entries, gateway notices, agent prompts and skills, shipped docs, specs and the feature map. Code names that carried the old name (components, CSS classes and tokens, test ids, i18n keys, tests, capture harnesses) follow it, with a ConductorMode / conductor_mode / conductor-mode stem so they never collide with the existing conductor agent-role names. Saved state keeps working: ids that never said ultra stay (crew_mode, dashboard.new_chat_crew_mode, mc-preview-crew-mode), the ultra_mode_only refusal code and the ultra root_kind trace value keep their spelling, and a transcript that stored the old agent-switch notice still renders as that notice. A new decision entry records the rename and maps the earlier entries to the new name; those entries stay as written. Hand-resolved backport; original author: Bolin Chen. (cherry picked from commit 176820c)
Ultra Mode only ever lived behind the default-off feature preview, so
no old spelling needs to keep working. Rename the four names the first
rename kept: the remote-dispatch refusal code is conductor_mode_only,
the config key is telemetry.traces_conductor_mode, and the trace
root_kind value and round trace kind are conductor_mode, with spans
named "conductor_mode round N" and "conductor_mode outcome ...".
Drop the compatibility code: the gateway notice no longer matches the
stored "Ultra Mode" agent-switch wording, and the comments that said a
value keeps its old spelling are gone. Title Case labels write the name
in lower case too ("Start New Chats in conductor mode"). A new decision
entry records both, superseding part of the entry that named the
feature.
Hand-resolved backport; original author: Bolin Chen.
(cherry picked from commit 36a018f)
…ename The rename changes the text of the two conductor agent specs and of _CONDUCTOR_SYSTEM_PROMPT. Main re-pinned the same pins in #36924, but release/0.13.0 builds those specs from its own prompts, so its digests differ from main's. Only kirocrew-conductor.json and kirocrew-ledger-conductor.json change; every other pinned spec keeps its bytes.
2b1cd20 to
cfa522f
Compare
Problem / Motivation
Goal: 0.13 users see the feature under its final name, conductor mode, so they do not meet "Ultra Mode" in 0.13 and a new name again in 0.14.
Backport onto
release/0.13.0(head22392729303, after #37514) of the two main PRs that renamed Ultra Mode to conductor mode. The rename is user-facing: UI strings in every locale, settings labels and their search entries, gateway notices, agent prompts and skills, shipped docs and specs. It also renames the code names, theconductor_mode_onlyrefusal code, thetelemetry.traces_conductor_modeconfig key and the trace values, as on main.176820cd7a36a018f0d7Both PRs passed review on main.
Why it matters
The feature ships in 0.13 behind its preview flag. Without the rename, 0.13 users who try it learn "Ultra Mode", then see every label, setting and notice change name in 0.14.
Not a goal
What changed (motivation → approach → change)
Two cherry-picks with
git cherry-pick -x, plus one test commit. Neither pick applies cleanly on 0.13: main renamed lines that 0.13 does not have yet, and 0.13 has lines main changed later. Every conflict was resolved the same way: keep 0.13's own text and apply the rename to it. Files main changed but 0.13 does not have (crew_log/projection_objects.py,test/test_warm_task_brief_welcome.py) stay absent. Two feature-map cells were set to main's exact wording, and the find_ui index (ui-index.generated.json,guidePlans.gen.ts) was regenerated withnpm run gen:ui. The decisions index gains the two rename entries.Hand-resolved hunks, each keeping both sides:
docs/feature-map/README.md, the notice rows (rename vs 0.13 text): theCrew dispatch noticeandRestart Trust noticerows take the rename; 0.13's own surrounding rows keep their text with the rename applied.docs/feature-map/README.md, the same block after fix: 0.13 backport batch 2, part 2 (Slack-reported KAS fixes, small completions) #37514: keeps fix: 0.13 backport batch 2, part 2 (Slack-reported KAS fixes, small completions) #37514's newAgent no-tools noticerow (from feat(kas): tell a chat when the user's agent lists no tools #36064) between the two renamed rows.docs/feature-map/README.md, two cells (Crew panel,Conductor mode switch) set to main's exact wording.docs/decisions/README.md: adds the two rename decision rows; 0.13's other rows stay.src/kiro_crew/docs/ui-index.generated.jsonandwebsite/src/uiLocations/guidePlans.gen.ts: regenerated withnpm run gen:uion this tree, not hand-merged.warm_task_brief_inputs.py,transcriptRenderers.tsx,chatRolesParity.contract.test.ts,test_mate_conductor_parity.py, the crew-mode, session and crew-log specs): 0.13's text with the rename applied.crew_log/projection_objects.pyandtest/test_warm_task_brief_welcome.pystay absent, as on 0.13.The check that the rename is complete: for every file under
src,website/src,testandconfig-baseline.json, the count of remainingultranames (ultra mode,UltraMode,ultra_,ultra-) equals main's count for that file.The test commit re-pins the conductor spec digests in
test/test_agent_refactor_spec_bytes.py(kirocrew-conductor.json,kirocrew-ledger-conductor.json) and the_CONDUCTOR_SYSTEM_PROMPThash intest/test_agent_refactor_contracts.py. Main re-pinned the same pins in #36924, but 0.13 builds those specs from its own prompts, so the values differ from main's. Every other pinned spec keeps its bytes.Backwards compatibility
Compatible for anyone who never turned on the Ultra Mode preview, which is off by default. A preview user who set
telemetry.traces_ultra_modemust usetelemetry.traces_conductor_mode, as on main (#37095).Tests
Release PRs get no backend test lanes in CI, so tests ran locally on this head (on top of #37514), one pytest process per batch of 20 files, no xdist:
test_kas_cli_floor.py::test_a_pinned_link_flipped_after_the_check_never_runs_unsandboxed) fails the same way on a cleanrelease/0.13.0.test_agent_refactor_spec_bytes.py,test_agent_refactor_contracts.py,test_mate_conductor_parity.py,test_work_ledger.py,test_agent_no_tools_notice.py, the find_ui tests andtest/tracing: 720 passed, 0 failed. The conductor spec digests re-pinned in the third commit still match after fix: 0.13 backport batch 2, part 2 (Slack-reported KAS fixes, small completions) #37514.Frontend on this head:
tsc -p tsconfig.app.jsonclean,scripts/check-i18n-keys.mjsOK,npm run gen:ui -- --checkclean, the 31 vitest files the rename touches pass. flake8 and black on every Python file in the diff: clean. The per-file count of remaining old names equals main's.The AI review lanes fail closed on this PR because its diff is over 1 MB. Both picks were reviewed on main.
Manual verification
N/A. A rename; the same change is on main.
Screenshots / video
Why no screenshot: a text-only rename of labels that main's PRs already reviewed; no layout changes.
Related Issues
Backport of #36924 and #37095.
Sibling sweep
Siblings: none -- the per-file count of remaining old names equals main's
Pattern harvest
Not generalizable: backport of an already-reviewed rename
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...). A backport keeps one commit per pick.Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license.