Skip to content

refactor: finish the conductor mode rename - #37095

Merged
bolichen97 merged 1 commit into
mainfrom
refactor/conductor-mode-drop-ultra-compat
Nov 14, 2026
Merged

bolichen97 merged 1 commit into
mainfrom
refactor/conductor-mode-drop-ultra-compat

Conversation

@bolichen97

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: No Ultra name is left in the product: the names the first rename kept for compatibility become conductor-mode names, and the compatibility code goes.

#36924 renamed Ultra Mode to conductor mode but kept a few old spellings so a saved value would keep working: the remote-dispatch refusal code ultra_mode_only, the config key telemetry.traces_ultra_mode, the trace value ultra, and a match for the stored "Ultra Mode" notice wording. Two Title Case labels also kept a capital M ("Start New Chats in Conductor Mode", "Traces: Conductor Mode Shape"). A maintainer ruling says none of that compatibility is needed, because Ultra Mode only ever lived behind the default-off feature preview.

Why it matters

Each kept spelling is a second name for one feature that a reader, a log query or a test has to know about, and the compatibility code exists only for installs that never had the feature outside a preview.

Not a goal

  • Changing what conductor mode does.
  • The retired reply-verbosity level ultra, effort values, Ultracode and jev-ultrafast: other features.
  • Editing an existing decision entry, an old RFC or CHANGELOG.md.

What changed (motivation → approach → change)

The last old spellings are gone, and the name reads the same way in every label.

Kept by #36924 Now
refusal code ultra_mode_only conductor_mode_only
config key telemetry.traces_ultra_mode telemetry.traces_conductor_mode
kirocrew.crew.root_kind value ultra conductor_mode (the other root kind stays mate)
round trace kind ultra, spans ultra round N / ultra outcome ... conductor_mode, conductor_mode round N / conductor_mode outcome ...
gateway notice matching the stored "Ultra Mode" wording removed, with its test case
"Start New Chats in Conductor Mode", "Traces: Conductor Mode Shape" "Start New Chats in conductor mode", "Traces: conductor mode Shape"

No new value collides with an existing one: conductor_mode is not a trace kind, run kind or root kind anywhere else (git grep).

The rename reaches the writers and readers of each name in one commit: src/kiro_crew/dashboard/handlers/remote_dispatch_routes.py and the goal-conductor skill in src/kiro_crew/builtin_skills; src/kiro_crew/config/sections_telemetry.py, config/loader_builders.py, src/kiro_crew/tracing/provider.py; src/kiro_crew/tracing/crew.py, tracing/work.py, src/kiro_crew/dashboard/trace_binding.py; website/src/pages/chat/gatewayNotice.ts. It also covers the label in website/src/i18n and its references in website/src/components, website/src/uiLocations and website/src/pages/chat/page/crewModeSwitch.ts; the docs in docs/system-specs, docs/guides, docs/feature-map and src/kiro_crew/docs; and the regenerated config-baseline.json, settings registry and UI index.

A new decision entry, docs/decisions/2026-11-13-conductor-mode-keeps-no-ultra-names.md, records the ruling and supersedes, in part, the entry #36924 added; that entry stays as written.

After this change git grep -i ultra finds only the unrelated items above, vendored code, old decision entries and their README rows, old RFCs, the task-spec archive and CHANGELOG.md. git grep 'Conductor Mode' finds nothing.

Backwards compatibility

Breaking: the refusal code, the config key and the trace values change with no fallback, and a transcript that stored the old "Ultra Mode" agent-switch notice shows it as a plain notice line. Ultra Mode only shipped behind the default-off preview, and the config key and trace names reached main on 2026-11-13. Writer sweep on origin/main: the refusal code is written by remote_dispatch_routes.py; the config key by sections_telemetry.py and loader_builders.py; the trace values by tracing/crew.py and tracing/work.py. Every reader is in this commit.

Reader: src/kiro_crew/builtin_skills/goal-conductor/SKILL.md:conductor_mode_only -- chat -- names the new code; test_a_plain_owner_chat_is_refused_conductor_mode_only
Reader: src/kiro_crew/tracing/provider.py:conductor_mode_shape -- chat -- reads telemetry.traces_conductor_mode; test/tracing/test_tracing_work.py
Reader: test/tracing/test_tracing_runs.py:root_kind -- chat -- expects conductor_mode
Reader: website/src/pages/chat/gatewayNotice.ts:agent_switch_blocked_crew_mode -- chat -- matches the "conductor mode" wording the gateway writes; gatewayNotice.test.tsx

Tests

The tracing, remote-dispatch and gateway-notice tests expect the new names; the settings tests expect the lower-case label.

Proof: not provable -- other: refactor (rename) PR with no behavior fix; tests pin names only

Local runs on this head: black (26.10.1), isort, flake8, mypy (4635 files), comment-history, brand name, file line cap, docs-lint, decision-ledger history, feature map, npm run i18n:check (diff-scoped), UI index check, tsc, semgrep (0 findings), the flake-pattern ratchet, test/tracing plus the remote-dispatch, config, goal-conductor and spec-golden tests (1733 passed), and 88 related frontend test files. One backend test, test/tracing/test_tracing_otlp.py::test_a_header_echoed_back_by_a_failure_reaches_no_log_record[library_error], fails the same way on a clean origin/main checkout.

Manual verification

I built the SPA before and after (gateway-free, dark, English) and checked the Settings labels: "Start New Chats in conductor mode" and the "Conductor mode" Feature Previews card fit on one line at 1400 px, and the card fits at 390 px.

Screenshots / video

Chat settings label before

Chat settings label after

Related Issues

Closes #37086
Part of #36869

Follows #36924.

Waiting for a maintainer: the new decision entry supersedes part of a recorded decision, so recorded-decisions-are-not-regressions blocks this PR by design (docs/decisions/README.md rule 3) until a maintainer restates the ruling on it; the merge approval of a maintainer.

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

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks. The rename reaches every in-tree writer and reader, and the author lists and accepts the breaks for names that only existed behind the preview.

Notes

  • The deleted comment said "a conductor on another machine matches on it". That is a reader outside this tree. A gateway on the old version will not recognise the new conductor_mode_only refusal code. The Breaking: line accepts this, but it would help to say in the PR that a mixed-version conductor setup is accepted.

[DESIGN-REVIEWED] 7f7663c

Description read

[DESCRIPTION-READ] 2e0938d8f5398b8769fe099f64e104cf4ebc391bef96772f6ed1ba79450460d3

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 14, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 — 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 complete rename. A maintainer ruling linked in the description (#37086) covers each reversed compatibility pin, and the source tree still runs nothing from the old spellings.

What this change ships

Inventory (8 items) — 7 justified

Intent: drop the last Ultra names and their compatibility code so conductor mode has one name everywhere. This is a FIX (cleanup of #36924's leftovers).

  1. The remote-dispatch refusal code ultra_mode_only becomes conductor_mode_only, with no fallback. — justified
  2. The config key telemetry.traces_ultra_mode becomes telemetry.traces_conductor_mode, and a saved old key is no longer read. — justified
  3. The trace attribute kirocrew.crew.root_kind now reports conductor_mode instead of ultra. — justified
  4. The round trace kind and the span names ultra round N / ultra outcome … become conductor_mode …. — justified
  5. A stored "Ultra Mode" agent-switch notice is no longer rewritten and shows as a plain line. — justified
  6. Two labels now write the name in lower case: "Start New Chats in conductor mode" and "Traces: conductor mode Shape". — justified
  7. A new decision entry supersedes, in part, the naming entry refactor: rename Ultra Mode to conductor mode #36924 added. — justified
  8. A grammar fix in a crewModeSwitch.ts comment ("an" → "a"). — rides along, undeclared

[FIRST-PRINCIPLES-REVIEWED] 7f7663c

Description read

[DESCRIPTION-READ] c657389ca46199b9e994e0dc487de5677efaeafefe0e24998f0f0a1fce00da70

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 14, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS

UX-level review of 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

UX-Verdict: PASS

Nothing blocks: the one dashboard label this PR changes is in shot-02, and the blind reader correctly saw it as the same setting under a new name.

Notes

  • The reader said the lowercase "Start New Chats in conductor mode" "looks uneven" next to Title Case headings. The new decision entry chose this on purpose, so it stays. Still, the setting appears to have no visible switch in the shot ("I can't see any switch, checkbox or button"), and that is worth checking in the uncropped panel.
  • No screenshot shows Traces: conductor mode Shape, and its help text puts the raw span name conductor_mode round N in front of the user. It looks like it only renders through the config schema, not a dashboard control, so it is not a gap.

[UX-REVIEWED] 7f7663c

Description read

[DESCRIPTION-READ] 48280c86f631ae7bde066f4da110fdc29fece9e363c283c3937f01eecf9fd563

That sha256 names everything this verdict read: the PR description, plus the 5 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 14, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 7f7663c

Verdict parsed from the review's SHA-scoped output markers for commit 7f7663cc47150f4e43fd93edca55a4dfa06d31c0.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 7f7663c

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 7f7663cc47150f4e43fd93edca55a4dfa06d31c0 false-positive: <reason, 20 to 500 characters>

@bolichen97
bolichen97 enabled auto-merge (squash) November 14, 2026 05:20
@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override gpt 989564d not-applicable: agent: owner-authorized (maintainer ruling): the blocking finding is the by-design recorded-decisions gate on the new superseding decision entry; the diff is reviewed locally by GPT and Opus, no open findings; evidence in the PR body

@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override fable 989564d not-applicable: agent: owner-authorized (maintainer ruling): the blocking finding is the by-design recorded-decisions gate on the new superseding decision entry; the diff is reviewed locally by GPT and Opus, no open findings; evidence in the PR body

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 recorded a not-applicable release (class not-applicable) for the gpt AI finding on 989564dcd44652232c876dfcb4515decd3aa23da.

agent: owner-authorized (maintainer ruling): the blocking finding is the by-design recorded-decisions gate on the new superseding decision entry; the diff is reviewed locally by GPT and Opus, no open findings; evidence in the PR body

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

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 recorded a not-applicable release (class not-applicable) for the fable AI finding on 989564dcd44652232c876dfcb4515decd3aa23da.

agent: owner-authorized (maintainer ruling): the blocking finding is the by-design recorded-decisions gate on the new superseding decision entry; the diff is reviewed locally by GPT and Opus, no open findings; evidence in the PR body

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

@bolichen97

Copy link
Copy Markdown
Contributor Author

CI red on test/test_search_docs.py::test_search_docs_ranks_the_page_the_query_is_about_first[...how do I run a task in the background] (Backend Tests (3.12, 6)) also fails on a clean main checkout; it is tracked in #37060 with the fix in #37063. This PR waits for that fix. CodeQL (python) failed on an API rate limit and Dependency License Gate fails repo-wide on the dependency-graph API; neither is about this diff.

@KiroCrewCaptain

KiroCrewCaptain commented Nov 14, 2026 •

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: Approve

Why: the decider estimates a 12% chance that this change introduces a defect, within the 12% limit for approving on its own, and every check passes.

This PR finishes the rename of Ultra Mode to conductor mode. The rename is complete: ultra_mode_only, telemetry.traces_ultra_mode, the ultra root and round kinds, the ultra round / ultra outcome span names and the "Ultra Mode" notice regex are gone from src/, website/ and test/. Every writer and reader is updated in the same commit: the config dataclass, loader_builders, both reads in provider.py (including the inputs the settings watcher compares), tracing/crew.py and tracing/work.py, the generated registry and UI index, the i18n strings and the docs. The refusal stays a refusal, so nothing fails open, and the loader keeps unknown config keys, so an old value is not deleted. The breaks have no fallback, but the PR declares them, they affect a preview feature only, and the new decision entry records them.

An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit 7f7663cc47 · 195 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@bolichen97
bolichen97 force-pushed the refactor/conductor-mode-drop-ultra-compat branch from 989564d to 3f2d718 Compare November 14, 2026 05:55
@bolichen97

Copy link
Copy Markdown
Contributor Author

Rebased onto main to pick up the main-red fix #37063 (02bad16 -> a67cbbb). The patch is unchanged (git range-diff reports it equal); CI re-running.

@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 14, 2026
@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override fable 3f2d718 not-applicable: agent: owner-authorized (maintainer ruling): the blocking finding is the by-design recorded-decisions gate on the new superseding decision entry; the diff is reviewed locally by GPT and Opus, no open findings; evidence in the PR body

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 recorded a not-applicable release (class not-applicable) for the fable AI finding on 3f2d7183e6166f4827a114a67d2e5970ff607b28.

agent: owner-authorized (maintainer ruling): the blocking finding is the by-design recorded-decisions gate on the new superseding decision entry; the diff is reviewed locally by GPT and Opus, no open findings; evidence in the PR body

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

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Nov 14, 2026
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.
@bolichen97
bolichen97 force-pushed the refactor/conductor-mode-drop-ultra-compat branch from 3f2d718 to 7f7663c Compare November 14, 2026 07:29
@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Nov 14, 2026

@KiroCrewCaptain KiroCrewCaptain left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On head 7f7663c this review checked each renamed name (the conductor_mode_only refusal code, the telemetry.traces_conductor_mode key, the root_kind and round kind values, the span names and the notice regex) against its writer and every in-tree reader. It also grepped the merge of this PR into main for leftover Ultra names and found none outside the Not-a-goal areas. It found no correctness or security defect, and no blocking rule broken in code. PR Readiness is green on this head.
Done well: each rename reaches its writer and all of its readers in one commit, the generated config baseline and settings registry are regenerated rather than edited by hand, and the decision change is a new superseding entry that leaves the earlier one as written.
Optional:

  • The Backwards compatibility text says the breaking values are main-only. In fact root_kind = "ultra", ultra_mode_only and the stored "Ultra Mode" notice shipped in 0.13.0 behind the default-off preview. Only the config key and the round and span names are main-only. The Breaking: line already names all four changes.
  • No test reads traces_conductor_mode from a config file, and the refusal test compares against the constant, not the literal. A config-driven conductor_mode_shape() test and a literal "conductor_mode_only" assert would pin both renames.

@bolichen97
bolichen97 merged commit 36a018f into main Nov 14, 2026
97 of 98 checks passed
@bolichen97
bolichen97 deleted the refactor/conductor-mode-drop-ultra-compat branch November 14, 2026 09:24
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 14, 2026
buluoray added a commit that referenced this pull request Nov 15, 2026
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)
buluoray added a commit that referenced this pull request Nov 15, 2026
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)
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.

Conductor mode: drop the Ultra spellings kept for compatibility

2 participants