Skip to content

chore(i18n): call a connected machine a crew in every user string - #18436

Merged
iamwhatever merged 1 commit into
mainfrom
chore/10481-crew-vocabulary
Oct 9, 2026
Merged

iamwhatever merged 1 commit into
mainfrom
chore/10481-crew-vocabulary

Conversation

@iamwhatever

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Every user-visible string calls another computer you connected a "crew", so Settings, the sidebar, the crew window and their notices no longer name it three different ways.

The same connected machine was called a "remote instance", a "crew", a "server badge", "the remote", "the remote host" and "this machine" depending on where you looked. #12697 moved most strings back to "crew". Eight English strings still used another noun.

Why it matters

A user who reads "remote crew" in Settings, "server badge" in Feature Previews and "That instance" in the Send-to menu can fairly think these are three different things. One noun makes the feature read as one concept. The owner picked crew for #10481.

Not a goal

  • Renaming code symbols, API fields, config keys, i18n key names or routes.
  • "instance" where it means an EC2 instance or an AWS Identity Center instance.
  • The Mochi app's multi-instance strings, which refactor(i18n): restore the Crew vocabulary for remote crews #12697 kept on purpose.
  • "remote" as an adjective on SSH form fields ("Remote port", "Remote user").
  • Docs and CHANGELOG history.

What changed (motivation → approach → change)

Eight English strings in website/src/i18n/locales/en.json still named the connected machine with another noun. Each now says "crew":

Where Before After
Feature Previews, remote crew sessions each marked with a server badge each marked with its crew's name
Send to crew, transcript-only hint That instance could not restore… That crew could not restore…
Crew pane, dashboard never loaded the remote dashboard never loaded the crew's dashboard never loaded
Settings → Remote Crew, kirocrew path hint on the remote … run on the remote host on the crew … run on the crew
Settings → Remote Crew, name placeholder Remote Host 1 Crew 1
Settings → Remote Crew, session-manager plugin reach the instance reach the crew
Settings → Remote Crew, hand-added row does not manage this machine does not manage this crew
Settings → Remote Crew, cloud-unverified row whether this machine has AWS resources whether this crew has AWS resources

All 11 translated catalogs carry the same change in each language's own word for crew. The rewritten strings follow each style guide's register (tu, du, ты, तुम, তুমি), which moves six strings off the formal form (bn, de, fr, hi, ru). en-XA comes from npm run i18n:pseudo. settingsRegistry.gen.ts and src/kiro_crew/docs/settings-registry.generated.json come from npm run gen:settings; each moves by the one Feature Previews line. Two panel tests that matched the old wording now match the new.

Backwards compatibility

Compatible: only display copy changes; every i18n key, API field and config key keeps its name.

Removes nothing: no key, field, route or record is removed or renamed.

Tests

  • InstancesPanel.test.tsx and RemoteCrewPanel.test.tsx match the new placeholder and row notes. Several of those assertions are not.toBeInTheDocument(), so they would have passed silently on the old wording; they now check the new text.
  • The 12 *Style tests with I18N_BASE_REF=origin/main: 117/117. They hold every changed value to its locale's register and punctuation rules (French U+202F before ; and :).
  • src/i18n, both panel tests and commandPalette: 74 files, 1212 tests pass. npm run i18n:check (parity, DNT, pseudolocale, changed-passthrough) passes, as do tsc -p tsconfig.app.json and eslint on the touched files.
  • test_settings_deeplink_registry.py and test_brand_name_gate.py: 91 passed. Re-running both generators after rebasing on main gives no diff.
  • No new test, so no determinism proof applies.

Manual verification

Rendered the real RemoteCrewPanel, CrewGroupSection and CrewChatWindow in a local Vite harness (api client stubbed in the page, English), once on origin/main and once on this branch.

Screenshots / video

Settings → Remote Crew, with the add form open. The row note, the name placeholder and the path hint change.

Before:

Settings Remote Crew before

After:

Settings Remote Crew after

Sidebar crew group and crew window header (no change: they already said crew)

Both PNGs are byte-identical between main and this branch. #12697 had already moved these two surfaces to "crew", and they are shown here as the consistency check.

Sidebar crew group

Crew window header

Related Issues

Closes #10481

Part of #6180

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

@iamwhatever
iamwhatever requested review from a team and KiroCrewCaptain as code owners October 9, 2026 09:57
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 9, 2026
@github-actions

github-actions Bot commented Oct 9, 2026

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Check that "Crew 1" reads as a machine placeholder in Settings → Remote Crew, since "crew" also names an agent team elsewhere in the product.

[DESIGN-REVIEWED] 3eb7bfc

Description read

[DESCRIPTION-READ] 2bdf6c9305494f209c1e42a10b45689d3696c878bda9a7a8e89fc8a574ae3b08

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 Oct 9, 2026

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f — 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/CONCERNS are advisory.

First-Principles-Verdict: PASS

Nothing to check.

What this change ships

Inventory (8 items) — 8 justified

Intent: make every user-visible string call a connected machine a "crew", so the feature reads as one concept. This is a FIX (copy consistency; #10481 records the owner's choice).

  1. The Feature Previews hint now says remote sessions are "marked with its crew's name", not "a server badge". ChatSidebar.tsx:6680 does render instance_name. — justified
  2. The Send-to transcript-only hint now says "That crew", not "That instance". — justified
  3. The crew pane's never-loaded notice now says "the crew's dashboard". — justified
  4. The kirocrew path hint in Settings → Remote Crew now says "on the crew", not "remote host". — justified
  5. The name field placeholder now says "Crew 1", not "Remote Host 1". — justified
  6. The session-manager plugin note now says "reach the crew". — justified
  7. The note on a row you added yourself now says "does not manage this crew". — justified
  8. The cloud-unverified row note now says "this crew", and the six rewritten bn/de/fr/hi/ru strings move to the informal register their style guides require. — justified

[FIRST-PRINCIPLES-REVIEWED] 3eb7bfc

Description read

[DESCRIPTION-READ] c795741641b1aa6774c1c79a8a467378fecfa755a5adb9b91a86a9a475acbc49

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 Oct 9, 2026

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — 🟡 CONCERNS

UX-level review of 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

UX-Verdict: CONCERNS

Next to the product name, "crew" for a machine reads oddly: "Kiro Crew does not manage this crew" and "run on the crew" made the blind reader stumble.

Watch

  • The product name and the noun now sit side by side. The doesnt_manage note renders "{{productName}} does not manage this crew", which shot-02 shows as "Kiro Crew does not manage this crew". unverified_cloud_note has the same shape: "cannot verify whether this crew has AWS resources".
    • The blind reader's words: "the app's name makes 'crew' sound like a team", and "run on the crew" "reads oddly. It treats the crew as a computer."
    • Severity is moderate. Every hand-added or cloud-unverified row shows this note, every time. It causes friction but not task failure.
    • Smallest fix: name the row instead of the noun, e.g. "{{productName}} does not manage devbox" (pass the crew's name in).
  • The path hint asks the user to run a shell command "on the crew". is_installed_somewhere_non_standard_on_the_remot now says "To find it, run on the crew: command -v kirocrew". A terminal command needs a host, and "crew" doesn't read as one to a newcomer.
    • Severity is low. Only users with a non-standard install read this hint, and once.
    • Smallest fix: "run this in a terminal on that crew's computer:".
  • Five changed notices appear in no screenshot: plugin_missing, unverified_cloud_note, transcript_only_hint, the crew pane's "dashboard never loaded" notice, and the Feature Previews line. They are copy-only changes to existing notices, not new controls, so I treated them as unverified rather than blocking. The same "Kiro Crew … this crew" collision applies to unverified_cloud_note.

[UX-REVIEWED] 3eb7bfc

Description read

[DESCRIPTION-READ] 534cec980950698084199bbd3a80c62101d4cbd829ab265e4106499b4b4bb481

That sha256 names everything this verdict read: the PR description, plus the 7 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 Oct 9, 2026

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 3eb7bfc

Verdict parsed from the review's SHA-scoped output markers for commit 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f: <one-sentence reason>

@github-actions

github-actions Bot commented Oct 9, 2026

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 3eb7bfc

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 3eb7bfce5ec8a8a0ad6b6527f63994b9f595137f: <one-sentence reason>

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Oct 9, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Severity of the doesnt_manage note span=01488767111c

rebutted (disproportional). The finding rates it friction, not task failure; the note keeps the noun the panel's own heading uses, so no reader has to map two words to one thing.
No code change: re-running the whole CI fleet on a green copy-only PR for an advisory wording note is out of proportion.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Severity of the path hint span=0ebdd9b4db41

rebutted (disproportional). The finding rates it low and read once by non-standard installs only; no CI round is spent on it.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Blind reader: crew sounds like a team span=17334512fb30

rebutted (owner ruling). Reading crew as a team is exactly the ambiguity #10481 settled: the owner picked crew as the one noun for the machine, and the Feature Previews line states "a remote crew is another machine you have connected".
#10481's owner fixed the noun: crew means another computer you connected. The copy uses that meaning on purpose.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Smallest fix: run this in a terminal on that crew's computer span=2c02083a4758

rebutted (disproportional). The suggestion would reintroduce a second noun ("computer") beside crew, the very mix #10481 removes; the current wording already names the host.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Smallest fix: name the row instead of the noun span=2d024ee1f583

rebutted (out of goal). Interpolating the crew's name adds a new {{name}} variable to doesnt_manage and unverified_cloud_note in 12 catalogs and a call-site change in RemoteCrewPanel.tsx; the goal is copy only, no new parameters.
The row's own name sits on the same row directly above the note, so the referent is already visible.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Path hint says run a command on the crew span=332a225880e3

rebutted (owner ruling). Under the owner's definition a crew is a computer, so "run on the crew" names a host; the hint sits in the SSH form whose host field is the crew's own host alias.
#10481's owner fixed the noun: crew means another computer you connected. The copy uses that meaning on purpose.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Product name and noun side by side in doesnt_manage span=8c413e9ae98e

rebutted (owner ruling). "Kiro Crew" is the product brand; "crew" is the owner-chosen noun for a connected machine, and Settings already pairs them ("Remote Crew", "Your crews").
#10481's owner fixed the noun: crew means another computer you connected. The copy uses that meaning on purpose.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • Five changed notices appear in no screenshot span=fd368623104e

rebutted (evidence in body). Each of the five is a one-noun swap shown verbatim in the PR body's before/after table; the reviewer itself treated them as copy-only changes to existing notices.
The three surfaces the issue names (Settings, sidebar crew group, crew window) are captured.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Self-review (crew-pr-reviewer, read-only): RECOMMEND APPROVE

Copy-only PR: the eight English strings that still called a connected machine an "instance", "machine", "remote host" or "server badge" now say "crew", with the same change in all 11 translated catalogs, a regenerated en-XA and both generated settings registries. No i18n key, API field or config key is renamed. I read the full diff. The en, de and zh-CN values match the table in the body, and the two embedded-newline strings were collapsed, which renders the same in JSX. The two panel tests now match the new wording, which matters because several of those not.toBeInTheDocument() assertions would otherwise pass for nothing. All checks are green. Design, First Principles, GPT and Opus pass, and every item in the one UX CONCERNS on head 3eb7bfc has its own disposition, resting on the owner's naming ruling. No change requests are open. Nit only: the name placeholder is "Crew 1" in English but "Meine/Mi/Ma/La mia/Minha crew 1" in five locales, which is harmless for a placeholder. — crew-pr-reviewer

@iamwhatever
iamwhatever enabled auto-merge (squash) October 9, 2026 16:34

@buluoray buluoray left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Approving at 3eb7bfc.

What this fixes for users: the connected machine was "remote instance", "server badge", "the remote host" and "this machine" depending on the screen; after this every user string calls it a crew, in each language's existing word for crew.

Decision check: there is no entry in docs/decisions/ on this naming, and the noun matches the accepted remote-crew-sidebar RFC and the maintainer's comment on #10481 ("a crew is another computer you connected"). No contradiction.

What I checked

  • All 13 catalogs (12 languages plus en-XA) change exactly the same 8 keys versus the merge base: no key added, removed or renamed, and no other value touched. {{productName}} and <0>…</0> tokens are identical before and after in every changed value.
  • Each locale uses the same noun its neighbouring crew strings already use (zh-CN Crew, ja クルー, ko 크루, ru команда, hi क्रू, bn ক্রু, Crew/crew elsewhere). French keeps U+202F before ; and :.
  • All 8 keys are read through i18nT at their call sites and none lives in en.manual.json, so editing en.json is the right source. No code identifier, API field, config key or route changes. The two generated settings registries differ only by the one Feature Previews line and match en.json.
  • No leftover old wording outside the catalogs.
  • Locally: the two panel tests plus the i18n parity/glossary/pseudolocale/changed-value/style suites, 324 tests, pass, and the i18n check passes. I did not run the Python tests, tsc or eslint locally; CI on this head is 90 success and 7 skipped, with no failures.

Non-blocking follow-ups

  1. The i18n gates do not catch a locale that keeps its old meaning after the English source changes: reverting the German name placeholder to the old "Remote-Host 1" left every i18n test and the i18n check green. This was already true before this PR. Worth an issue so a changed English value flags any locale whose value did not change with it.
  2. The name placeholder reads "Crew 1" in English and most locales, but "My crew 1" in de/fr/es/pt/it. It is harmless for a placeholder, and only noted for consistency.
  3. Two Remote Crew hints still say "remote gateway". They name the gateway process on the crew rather than the crew, so they are outside this PR's scope, but could be polished later.

@iamwhatever
iamwhatever merged commit 10ca851 into main Oct 9, 2026
147 of 148 checks passed
@iamwhatever
iamwhatever deleted the chore/10481-crew-vocabulary branch October 9, 2026 19:17
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 9, 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.

Unify remote-instance vocabulary across session surfaces

2 participants