claude signal.

browser · author + co-author · 2026-10-07T03:41:38.000Z

An administrator's password card can no longer be edited while the change runs (#427)

An administrator's password card can no longer be edited while the change runs (#427) Closes #381 While an administrator's password change runs, the Your password card now goes quiet: its fields, Show toggles and button are disabled until the server answers, so nothing can be typed and no second change can start. This matches the vault owner's Change password card. <details> <summary>Technical</summary> - `passwordCard()` in `solvent/static/js/page-admin.js` gets the `quiet()` that `changePasswordCard()` in `view-settings.js` has. It disables every input and button in the fields while working and keeps the button disabled, and re-enables them after the answer. - `spec/features/admin-invites.md`: criterion 58. - Tests: a held change in `tests/browser/parts/admin.mjs`, and the reviewer's `tests/browser/parts/admin-review-working.mjs`. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

LiamBMX/joshymc-pluginindependent

feat(staff): /sj toggle for private silent join/leave notifications (#1063) (#1064)

interface · author + co-author · 2026-10-07T03:39:46.000Z

feat(staff): /sj toggle for private silent join/leave notifications (#1063) (#1064) When a vanished or joshymc.silentjoin player joins or leaves, other online joshymc.silentjoin staff with /sj on (default) get a private [Silent Join]/[Silent Leave] line. The preference is a hidden, cached player setting, so it persists by UUID across reconnects and restarts. Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Opus 5.5

mazyu36/aws-oss-updateindependent

feat: add bedrock-agentcore-sdk-python v1.24.1 release article (#92)

workflow · author + co-author · 2026-10-07T03:25:56.000Z

feat: add bedrock-agentcore-sdk-python v1.24.1 release article (#92) Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Opus 4.7

browser · author + co-author · 2026-10-07T03:14:46.000Z

Screen readers name each value field on Update values by its holding (#426)

Screen readers name each value field on Update values by its holding (#426) Closes #197 A screen reader now names each value field on Update values by its holding, such as "Mortgage value", the way the rate fields already read "Gold, troy ounce rate". <details> - `solvent/static/js/view-sweep.js`: the sweep row's quantity input carries `aria-label` of the holding's name plus " value", set through `el`, which uses `setAttribute`, so the decrypted name never renders as markup. - `spec/features/record-snapshot.md`: Update values, Layout, states the field's name. New criterion 105. - Tests: a check in `tests/browser/parts/update-values.mjs`, and the reviewer's `tests/browser/parts/update-values-review-field-names.mjs`, which reads the computed accessible name on today's sweep, a new and a reopened recording (archived holding included), after Record, at phone width, and for a name holding markup. Both fail on `master`. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

browser · author + co-author · 2026-10-07T02:46:53.000Z

Lists of names in a sentence read "A, B and C" (#425)

Lists of names in a sentence read "A, B and C" (#425) Closes #196 The delete confirmation for a recording now names the units it moves as "USD, XAU-ozt and XAG-ozt" instead of chaining them with "and". Every other list of names in a sentence reads the same way: the archived holdings and what a partial delete keeps on a recording, the units whose prices were not updated when recording or updating values, and the notice after leaving with unsaved changes. <details> <summary>Technical</summary> - `listed()` in `solvent/static/js/dom.js` joins names as "A", "A and B", "A, B and C". `view-recording.js`, `view-forms.js`, `view-holding-form.js`, `view-sweep.js` and `app.js` use it in place of `join(', ')` and `join(' and ')`. - The verb already agreed with the count on `master` ("1 holding ... moves"). - `spec/design-system.md`, Figures, gains the rule. `spec/features/record-snapshot.md` updates the leave notice copy and adds criterion 104. - Tests: `tests/browser/parts/recording-detail-units.mjs`, and reviewer's `*-review-names.mjs` parts for Recording detail, Snapshot entry, Update values and Account detail. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

workflow · author + co-author · 2026-10-07T02:02:22.000+03:00

Port the headless watch-command and widget scripts to Swift Testing and delete them (#40)

Port the headless watch-command and widget scripts to Swift Testing and delete them (#40) * Port the ten headless watch-command and widget scripts to Swift Testing and delete them Part 2 of #6. The scripts each compiled a slice of the app against stubs, and three of those stubs behaved differently from the app: acceptWatchWorkout always linked, discardOrphanWorkout always succeeded, and publishHeadless recorded every call. The new tests run the real code in the hosted app. WatchCommandCenter gains one seam: its Health save, backfill, orphan discard and phone-workout note go through `health`, each the real call by default, and the Health task it starts is kept in `healthFollowUp` so a test can await it rather than sleep. HealthKitService has a private init and no store in a test host, so without it nothing could see what the headless path asked of Health. The rig the headless tests drive the center with moves to GymTrackTests/Support so later parts can use it. The widget tests take their day from TestClock, so they no longer fail between midnight and 00:18 or on daylight-saving days. Co-authored-by: Marwan Abudaif <126610607+MarwanMoh11@users.noreply.github.com> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Port the three late-log repairs the first pass missed late-wrist-log-repairs also checked that a set the phone took back stays taken back through the wrist's Finish, before and after the phone's own Finish, and that a late log, its undo and a re-log land in either order. Those blocks were left out of the port; they are tests now, so every assertion of the deleted script has a home. Co-authored-by: Marwan Abudaif <126610607+MarwanMoh11@users.noreply.github.com> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Marwan Abudaif <126610607+MarwanMoh11@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Opus 5.5, Claude Opus 5.5, claude[bot], Claude Opus 5.5

rpvessoni-sys/Commodities-Radarindependent

leitura 2026-10-07 (#107)

implementation · author + co-author · 2026-10-07T01:55:43.000Z

leitura 2026-10-07 (#107) Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Sonnet 5.5

testing · author + co-author · 2026-10-07T01:47:43.000Z

A recording no longer says silver or m2 will be filled in after Update (#424)

A recording no longer says silver or m2 will be filled in after Update (#424) Closes #195 A recording no longer says a unit nobody publishes, such as silver or m2, "is filled in after Update". It names a unit as empty only when its price source did not answer for that date. <details> <summary>Technical</summary> - `view-recording.js`: the missing-price lines keep only units where `vault.quotable(unit, date)` holds, and none on a date that is not a recorded day, since a misdated recording has no Update and no source prices it. - `record-snapshot.md`, Recording detail, States: a unit with no rate source at the date is not named. New criterion 103 (blind). - Tests: a check in `recording-detail.mjs`, and the reviewer's `recording-detail-review-unpriced.mjs`. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

interface · author + co-author · 2026-10-07T01:46:24.000+03:00

Give the Library's first open a minute in the UI tests, so a busy CI simulator stops failing them (#38)

Give the Library's first open a minute in the UI tests, so a busy CI simulator stops failing them (#38) Two pull requests that never touched the Library went red in ui-tests because its title, and in the other run its exercise count, missed the ten-second wait after the tab was tapped. Measured on an idle runner, the app spends about half a second building the screen (the full-library search is under ten milliseconds of that), and the screen appears about four and a half seconds after the tap, up to a second behind the other tabs. The failed runs had a simulator that took 45 seconds to launch and 36 for one accessibility snapshot, so the slowest tab was the one to run out of time. The three waits for the first thing on the Library now allow a minute, which outlasts a snapshot that slow. Every assertion is unchanged, and a Library that never opens still fails. Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Marwan Abudaif <126610607+MarwanMoh11@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Opus 5.5

context · author + co-author · 2026-10-07T01:29:38.000+03:00

Keep the unit-test host's window empty, so RootView's launch work never resets what a test set up (#37)

Keep the unit-test host's window empty, so RootView's launch work never resets what a test set up (#37) A unit-test host already skipped configureServices so the tests own the singletons, but its window still showed RootView. That view's launch work then ran against the host's empty in-memory store whenever the window appeared: setCustom([]) and the hidden list on the shared catalog, the watch link's screen handler, the watch bridge, the widgets and Health. Landing after a test had set one of those up would change what it saw. The host now shows nothing in its window. A real launch and a UI-test launch still show RootView, and the UI tests keep proving it. Closes #8 Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Marwan Abudaif <126610607+MarwanMoh11@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Opus 5.5

workflow · author + co-author · 2026-10-07T01:26:44.000+03:00

Send a Start from the watch live instead of behind a queued request for the mirror (#31)

Send a Start from the watch live instead of behind a queued request for the mirror (#31) The wrist queued a request for the mirror whenever it could not send it live: at launch, on a wrist raise, on a return to range, and whenever a live send failed. While anything sat in that queue, every later command went into it too, a Start included, and the system delivers the queue when it chooses. So a Start tapped after any of those waited there instead of going live: the watch sat on "Starting" and gave up, and the phone began the workout whenever the queue finally arrived. The request the idle screen sends three seconds after a Start, to recover a lost reply, was queued behind it the same way, so it never recovered anything. A request for the mirror now goes live or not at all, as a live heart-rate reading already did, and holds nothing up. A Start waits in the queue only behind the end of the last workout, since ahead of it the phone would answer with the session the wrist has just closed. Everything else keeps the order it was tapped in. The rule lives in WatchCommandRouting so it can be tested without a paired phone. Closes #20 Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Marwan Abudaif <126610607+MarwanMoh11@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Opus 5.5

browser · author + co-author · 2026-10-07T01:21:55.000Z

A typed price on a line nobody publishes a price for no longer says there is no price (#421)

A typed price on a line nobody publishes a price for no longer says there is no price (#421) Closes #193 When you type a price into a rate line for a unit nobody publishes a price for, the line now reads "Typed by you" and only says why the price is yours to set. It no longer also says "No price for m2 yet.", how old the last price is, or asks for a price. The same goes for a line dated before a unit's published prices begin. <details> <summary>Technical</summary> - `solvent/static/js/view-sweep.js`, `line.describe`: in the `early` and `!quotable` branches, a changed line takes precedence and shows only `earlyCopy()` or `ownCopy()`. - `spec/features/record-snapshot.md`, Screens: a statement for a typed price on a line with no rate source at this date. Criteria 16 and 17 cover it. - Tests: a new check in `tests/browser/parts/update-values.mjs`, and the reviewer's `tests/browser/parts/update-values-review-typed.mjs`. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

browser · author + co-author · 2026-10-07T00:53:47.000Z

Screen readers name the Password and Confirm password fields when registering (#420)

Screen readers name the Password and Confirm password fields when registering (#420) Closes #191 On both registration forms, a screen reader now names the Password and Confirm password fields by the label beside them. Username and Main currency were already named. <details> <summary>Technical</summary> - `solvent/static/js/register-form.js`: the two password inputs get ids `register-password` and `register-confirm`, and their labels point at them with `for`. - `spec/features/register.md`: new acceptance criterion 51. - Tests: a check in `tests/browser/parts/register.mjs` that each field's label names it, and the reviewer's `tests/browser/parts/register-review-names.mjs`, which reads the names from Chrome's accessibility tree. Both fail on `master` for the two password fields. - Main currency got its label in #402. The same unlinked labels on the administrator's and Settings screens are #388 and #385. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

electron/electronindependent

test: run WebContentsView focusOnNavigation and shell.openExternal specs serially (#54662)

browser · author + co-author · 2026-10-07T00:48:41.000Z

test: run WebContentsView focusOnNavigation and shell.openExternal specs serially (#54662) * test: wait for focus event in WebContentsView focusOnNavigation spec The WebContentsView focusOnNavigation block was not tagged serial, unlike the equivalent block in api-web-contents.spec.ts. On macOS another parallel worker (notably the shell.openExternal specs, which bring the default browser to the front) can take app activation away mid-test. The window then stops being key, ReadyToCommitNavigation skips focusing because the owner window is inactive, and isFocused() returns false. Run the block in the serial phase, and wait for the webContents 'focus' event after navigating instead of checking isFocused() as soon as loadURL resolves. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X1mwQWDtyUGtjT3hFLoqms * test: run shell.openExternal specs serially On macOS the shell.openExternal specs launch the default browser and bring it to the front. While they ran in the parallel phase, that took app activation away from focus-sensitive specs in other workers (for example the WebContentsView focusOnNavigation spec on macOS x64). Run them in the serial phase so nothing else is running when the browser activates. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X1mwQWDtyUGtjT3hFLoqms --------- Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Opus 5.5 (1M context), Claude Opus 5.5 (1M context), Claude

context · author + co-author · 2026-10-07T00:29:50.000Z

Product description says a price's source is shown on its recording, not on a holding's list of values (#419)

Product description says a price's source is shown on its recording, not on a holding's list of values (#419) Closes #188 Nothing in the app changes. Two descriptions of the product said a holding's own list of values shows the price on each date and where it came from. The app shows the value, its figure in your main currency and a link to that day's recording, where the price and its source are. Both descriptions now say so. <details> <summary>Technical part</summary> - `spec/features/rate-lookup.md`, "Where you meet a price": drops the holding's history, which shows converted figures but never a price or its source. - `spec/features/record-snapshot.md`, the "No screen of nothing but prices" bullet: provenance points to Recording detail instead of the holding's page. - Both now agree with `spec/features/manage-accounts.md`, Account detail ("No Rate, Source or price-date column"). - No code, test or acceptance criterion changes. `reviewer` found nothing. </details> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

Comfy-Org/ComfyUI_frontendindependent

feat(website): add CTAs and footer link to Nano Banana 2.1 page (#20370)

browser · author + co-author · 2026-10-07T00:27:26.000Z

feat(website): add CTAs and footer link to Nano Banana 2.1 page (#20370)  _Requested by **June** · [Slack thread](https://comfy-organization.slack.com/archives/C0B9ZNZFZB8/p1791330664897469)_ ## Summary **Before:** The /nano-banana page showed the intro copy and badges with no buttons under it, so a visitor had no direct way to run the model or open the guide. The footer Models column had no Nano Banana 2.1 entry. **After:** Two buttons sit below the intro copy: RUN NANO BANANA 2.1 (primary, opens Comfy Cloud with the Nano Banana 2.1 text-to-image template) and READ THE GUIDE (secondary, opens the Comfy Router guide for Nano Banana 2.1 in the docs). Both open in a new tab. The footer Models column now lists Nano Banana 2.1 and links to /nano-banana. How: the page already supports hero buttons (same as ChatGPT Images 2.5), so this sets `primaryCta` and `secondaryCta` on the Nano Banana hero data and adds `en` and `zh-CN` labels. The run link carries the same utm params as the page's existing Cloud link. The footer entry reuses the existing `routes.nanoBanana` route with a new `footer.nanoBanana` string in `en` and `zh-CN`. ## Verification - Localization: `en` and `zh-CN` strings added for both buttons and the footer link. The `/zh-CN/nano-banana` page already exists, so no new page twin was needed. The zh-CN button labels are "运行 Nano Banana 2.1" and "查看教程", copied from the ChatGPT Images 2.5 page wording; worth a native check. - Tests: `vitest run` on `src/templates/model-launch` and `src/components/common` (219 passed); `pnpm typecheck:website`, `pnpm --filter @comfyorg/website format:check` and eslint on the changed files are clean. - Not run: e2e and screenshots (no browser in this environment); see the deploy preview for the visual check. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01BBFcxj8fzUA7Dbvrnm69RP --- _Generated by [Claude Code](https://claude.ai/code/session_01BBFcxj8fzUA7Dbvrnm69RP)_ Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

browser · author + co-author · 2026-10-07T00:03:27.000Z

A used invite shows the day it was used and links its username to Accounts (#418)

A used invite shows the day it was used and links its username to Accounts (#418) Closes #187 A used invite now says who used it and on which day, for example "Used sarah on 14.03.2026". Clicking the name opens Accounts at that person's row. When the account has since been removed, the row reads "account removed" with the day, and nothing links. <details> <summary>Technical</summary> - `solvent/static/js/page-admin.js`: the Status cell reads chip, username, " on " and the day in the Created column's format. The username is a `link-button`, which switches to Accounts and passes the username to `accounts()`. The matching row's username cell gets `tabindex="-1"`, is scrolled into view and focused, by comparing fetched data, never a selector built from the username. An account removed while Invites was open is absent, and nothing is focused. - `solvent/static/css/tokens.css`: the focused username cell draws the petrol-600 focus ring, and the link button keeps a 44px target at phone width. - `spec/features/admin-invites.md`: the Status bullet states the day, the link and the focus, and that the username starts the next line where it does not fit beside the chip (at 320 and 901px the space beside the chip is 79px). Criterion 57 is new, and 51, 53 and 57 name the reviewer's part. - Tests: `tests/browser/parts/admin.mjs`, `tests/browser/parts/admin-phone.mjs` (a username button may split, as the page lets a username break anywhere), and the reviewer's `tests/browser/parts/admin-review-used.mjs`. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

Madrasatul-Muslimeen/Madrasatul-Muslimeen.github.ioindependent

Note pictures (#620): suite reports a refused insert as a check, not a crash

interface · author + co-author · 2026-10-06T23:56:32.000Z

Note pictures (#620): suite reports a refused insert as a check, not a crash Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Sonnet 5.5

Madrasatul-Muslimeen/Madrasatul-Muslimeen.github.ioindependent

Note pictures (#620): Storage Rules candidate, browser suite, owner-only fetch

browser · author + co-author · 2026-10-06T23:45:48.000Z

Note pictures (#620): Storage Rules candidate, browser suite, owner-only fetch Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Sonnet 5.5

Madrasatul-Muslimeen/Madrasatul-Muslimeen.github.ioindependent

Note pictures (#620): resize + upload module, Storage stub, 🖼 button, pictures shown when a Note is read

implementation · author + co-author · 2026-10-06T23:41:51.000Z

Note pictures (#620): resize + upload module, Storage stub, 🖼 button, pictures shown when a Note is read Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Sonnet 5.5

Madrasatul-Muslimeen/Madrasatul-Muslimeen.github.ioindependent

Note pictures (#620): sanitiser keeps data-mmsa-image paths only; path + step-down module; pure checks

implementation · author + co-author · 2026-10-06T23:39:39.000Z

Note pictures (#620): sanitiser keeps data-mmsa-image paths only; path + step-down module; pure checks Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Sonnet 5.5

electron/node-abiindependent

ci: set cache-mode on release workflows (#306)

workflow · author + co-author · 2026-10-06T23:25:53.000+01:00

ci: set cache-mode on release workflows (#306) Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/node-minidumpindependent

ci: set cache-mode on release workflows (#163)

workflow · author + co-author · 2026-10-06T23:25:41.000+01:00

ci: set cache-mode on release workflows (#163) Add top-level `cache-mode: none` to release.yml so the release run (and the test/build workflows it calls via `uses:`) never reads from or writes to the GitHub Actions cache. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/github-app-authindependent

ci: set cache-mode on release workflows (#66)

workflow · author + co-author · 2026-10-06T23:25:29.000+01:00

ci: set cache-mode on release workflows (#66) Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/symbolicate-macindependent

ci: set cache-mode on release workflows (#121)

workflow · author + co-author · 2026-10-06T23:25:21.000+01:00

ci: set cache-mode on release workflows (#121) Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/chromedriverindependent

ci: set cache-mode on release workflows (#193)

workflow · author + co-author · 2026-10-06T23:25:15.000+01:00

ci: set cache-mode on release workflows (#193) Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/devtronindependent

ci: set cache-mode on release workflows (#378)

workflow · author + co-author · 2026-10-06T23:24:59.000+01:00

ci: set cache-mode on release workflows (#378) Add top-level `cache-mode: none` to release.yml so release runs never read from or write to the Actions cache, regardless of per-step settings. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/rebuildindependent

ci: set cache-mode on release workflows (#1307)

workflow · author + co-author · 2026-10-06T23:24:57.000+01:00

ci: set cache-mode on release workflows (#1307) Add top-level `cache-mode: none` to release.yml and docs.yml so the release and docs-publish paths never read from or write to the Actions cache, regardless of per-step settings. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/tropindependent

feat: request original approvers on manual backports of claude[bot] PRs (#437)

context · author + co-author · 2026-10-06T23:23:43.000+01:00

feat: request original approvers on manual backports of claude[bot] PRs (#437) Manual backports request review from the original PR's author and the optional default review team. When claude[bot] opens both the original PR and its backport, nobody is requested. When the original PR's author is claude[bot], request the users whose latest review on it is an approval instead, skipping bots and the backport's author, and keeping the existing write-access check. Claude-Session: https://claude.ai/code/session_01ViXXtuJhgzS1RHhwnk26N5 Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/rebuildindependent

fix: include target platform in .forge-meta rebuild marker (#1319)

context · author + co-author · 2026-10-06T23:23:11.000+01:00

fix: include target platform in .forge-meta rebuild marker (#1319) The `.forge-meta` marker only recorded `${arch}--${ABI}`, so once a module had been rebuilt for one platform, a later run targeting a different platform (via the `platform` option) in the same node_modules would skip it and keep the previous platform's binary. Record the platform as well. Existing markers no longer match, so each module is rebuilt once after upgrading. Refs: electron-userland/electron-builder#10031 Claude-Session: https://claude.ai/code/session_01NEnGhaCaJpaCKqnpBqpWJs Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

browser · author + co-author · 2026-10-06T23:04:02.000Z

Adding a unit shows Rate lookup, held at Entered by hand with the reason (#417)

Adding a unit shows Rate lookup, held at Entered by hand with the reason (#417) Closes #186 Adding a unit now shows the Rate lookup setting. It is held at "Entered by hand", with the reason beside it, because every unit a price source serves is already in the list. A unit added without a source, such as the currency XTS, stays on "Entered by hand" in the unit list and cannot be switched to automatic. <details> <summary>Technical part</summary> - `solvent/static/js/page-admin.js`: the Add a unit form gets a disabled Rate lookup `select` at `false` with the no-source hint. The reason text is one constant shared with the table rows. The `POST` still sends `lookup: false`. - The server side, `hasAdapter` from the registry and `lookup: true` refused without an adapter, already held on `master` since #350. The reviewer's tests now pin it for added units. - `spec/features/admin-invites.md`: the form's Rate lookup line and criterion 56. - Tests: `tests/browser/parts/admin.mjs`, `tests/browser/parts/admin-review-unit-lookup.mjs`, four tests in `tests/test_review_rate_lookup.py`. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

browser · author + co-author · 2026-10-06T22:31:41.000Z

Adding a metal is refused unless its code names a weight (#416)

Adding a metal is refused unless its code names a weight (#416) Closes #185 An administrator can no longer add a metal whose code names no weight. The add form refuses it and shows the expected shape: `<code>-ozt` or `<code>-g`, such as `XAU-ozt`. A metal stored without a weight moves to Retired and cannot be restored. Holdings already measured in it keep their unit and their prices. <details> <summary>Technical</summary> - `solvent/rates.py`: `POST /api/admin/symbols` holds a metal to `METAL_PATTERN`, `^(?=.{1,16}$)[A-Z0-9][A-Z0-9._]*-(ozt|g)$`. Both patterns now use `fullmatch`, because `re.match` with `$` admits a trailing newline. `PATCH` with `retired: false` on a weightless metal is a Bad Request. - `solvent/db.py`: every start retires the metal rows that fail `METAL_PATTERN`. - `solvent/static/js/page-admin.js`: refusal copy for a metal names the shape. A retired weightless metal has Restore disabled, with its reason beside it. - Spec: `rate-lookup.md` (Maintaining the table, Seeded symbols, SSRF and egress hardening, criteria 81 and 82) and `admin-invites.md` (Units, its error states, criteria 54 and 55). - Tests: `tests/test_rates.py`, `tests/browser/parts/admin.mjs`, and the reviewer's `tests/test_review_rate_lookup.py` and `tests/browser/parts/admin-review-units.mjs`. - Filed outside this change: #414, #415. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

electron/windows-installerindependent

ci: set cache-mode on release workflows (#616)

workflow · author + co-author · 2026-10-06T22:25:44.000Z

ci: set cache-mode on release workflows (#616) Add top-level `cache-mode: none` to release.yml so the release run (and the test.yml jobs it calls) never reads from or writes to the Actions cache, regardless of per-step cache settings. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/mksnapshotindependent

ci: set cache-mode on release workflows (#179)

workflow · author + co-author · 2026-10-06T22:25:33.000Z

ci: set cache-mode on release workflows (#179) Add top-level `cache-mode: none` to release.yml so release runs never read from or write to the Actions cache, regardless of per-step settings. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/typescript-definitionsindependent

ci: set cache-mode on release workflows (#357)

workflow · author + co-author · 2026-10-06T22:25:32.000Z

ci: set cache-mode on release workflows (#357) Add top-level `cache-mode: none` to release.yml so the release run (and the test.yml jobs it calls) never reads from or writes to the Actions cache, regardless of per-step cache settings. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/windows-signindependent

ci: set cache-mode on release workflows (#124)

workflow · author + co-author · 2026-10-06T22:25:24.000Z

ci: set cache-mode on release workflows (#124)  _Requested by **David Sanders** · [Slack thread](https://electronhq.slack.com/archives/C0BERTA8EKW/p1789070712587299)_ **Before:** The `Release` workflow (push to `main`) and the `Publish documentation` workflow (tag push) had no workflow-level cache policy. Every job in them, including the `test` job that `Release` calls via `uses: ./.github/workflows/test.yml`, could read from and write to the GitHub Actions cache if any step asked for it; the only thing keeping them cache-free was the per-step `package-manager-cache: false` setting on `actions/setup-node`. **After:** Both workflows declare `cache-mode: none` at the top level, so no job in a release or docs-publish run can restore or save a cache entry regardless of per-step settings. Because `cache-mode` on the caller also caps any reusable workflow it calls, the `test.yml` jobs run cache-free when invoked from `Release` (they keep their normal caching when `test.yml` runs on its own for pull requests). The existing `package-manager-cache: false` lines are left in place. This uses GitHub's new workflow-level `cache-mode` key, which lets release paths be made cache-free by policy rather than by convention: https://github.blog/changelog/2026-09-10-control-github-actions-cache-access-with-cache-mode/ and https://docs.github.com/actions/reference/workflows-and-actions/workflow-syntax#cache-mode. **How:** Adds a commented `cache-mode: none` line after the top-level `permissions:` block in `.github/workflows/release.yml` and `.github/workflows/docs.yml`, preserving the files' existing CRLF line endings. No other lines change; `test.yml` is untouched so PR CI keeps its cache. Note that PR CI does not exercise the release or docs workflows, so the first release (and first tag push) after merge is the real test. If a step is denied cache access, `actions/cache` and the `@actions/cache` toolkit log a message and continue (a skipped restore behaves as a cache miss, a skipped save is a no-op) rather than failing the job. Local validation: the files parse as YAML and zizmor 1.30.1 reports no new findings; actionlint 1.7.12 flags `cache-mode` as an unexpected workflow key because its schema predates the feature (this repo does not run actionlint in CI). 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd --- _Generated by [Claude Code](https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd)_ Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/osx-signindependent

ci: set cache-mode on release workflows (#449)

workflow · author + co-author · 2026-10-06T22:25:19.000Z

ci: set cache-mode on release workflows (#449) Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/fusesindependent

ci: set cache-mode on release workflows (#136)

workflow · author + co-author · 2026-10-06T22:25:16.000Z

ci: set cache-mode on release workflows (#136) Add the workflow-level `cache-mode: none` key so release and docs-publish runs never read from or write to the Actions cache. Claude-Session: https://claude.ai/code/session_01MYe3WCXtgiqomNuVQdi4Yd Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

electron/build-toolsindependent

fix: include clang-tidy in siso clang toolchain subtree (#910)

context · author + co-author · 2026-10-06T22:23:28.000Z

fix: include clang-tidy in siso clang toolchain subtree (#910) siso uploads the third_party/llvm-build/Release+Asserts:headers filegroup as a precomputed subtree and skips any other input under that directory. The electron/clang-tidy rule lists third_party/llvm-build/Release+Asserts/bin/clang-tidy as an explicit input, which only reached the RBE workers because Chromium's glob included bin/clang-*. https://crrev.com/c/8471431 narrowed that glob to bin/clang-cl, bin/clang-cl.exe and bin/clang-[0-9]*, so clang-tidy is now dropped and the remote tidy steps fail with exit 127 ("clang-tidy: No such file or directory"). Add bin/clang-tidy back to the filegroup's includes. This is a no-op on older Chromium, where bin/clang-* already matches it. Claude-Session: https://claude.ai/code/session_01V38FjDEdZgMvFpf5Fmp65m Co-authored-by: Claude <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude

Madrasatul-Muslimeen/Madrasatul-Muslimeen.github.ioindependent

Link previews: put the menu item last so existing menu positions do not move (#619)

interface · author + co-author · 2026-10-06T22:05:47.000Z

Link previews: put the menu item last so existing menu positions do not move (#619) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author Claude Sonnet 5.5

testing · author + co-author · 2026-10-06T21:49:38.000Z

Typing a price where none is stored saves without a confirmation, and the confirmation names only prices that move a holding (#413)

Typing a price where none is stored saves without a confirmation, and the confirmation names only prices that move a holding (#413) Closes #184 Typing a price where none is stored for that date no longer asks first. This covers a price typed on the value form over an empty line, a proposal or an estimate. When you change a stored price on a recording you reopened, the confirmation leaves out any unit whose change moves no holding on that date. When nothing is left to name, it does not open. It no longer says "moves 0 holdings" or that your net worth changes when it does not. <details> <summary>Technical</summary> - `view-forms.js`: the form's price confirmation is gone. Its lines never hold a stored entry, so every save there fills in a missing price, as the sweep at a new date does. - `view-sweep.js`: `rateChangeCopy` drops an update whose `holdingsIn` count is 0. `saveRates` saves without a dialog when the copy is empty. A clear always confirms, because it moves the neighbors' stretch and can leave a unit unpriced. - Spec: `record-snapshot.md` (The prices line, Changing or clearing a rate says what it moves, criterion 30, new criterion 102) and `record-rate.md` (The rate-lines save, Editing a captured rate). - Tests: `snapshot-entry.mjs` (a proposal and an estimate changed on the form ask nothing), a `tests/client/run.mjs` check, and the reviewer's `snapshot-entry-review-first-price.mjs` and `update-values-review-first-price.mjs`. - The sweep's half of the report (filling an empty line on a reopened recording) was already fixed by #325. </details> --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>

author claude[bot] · committer web-flow · co-author claude[bot]

bamr87/lifehacker.devindependent

content(doc): trim over-length SEO description in two-locks-and-one-is-a-decoy (#688)

infra · author + co-author · 2026-10-06T21:49:02.000-06:00

content(doc): trim over-length SEO description in two-locks-and-one-is-a-decoy (#688) The `description:` frontmatter was 227 chars; lint_frontmatter.rb's SEO soft cap is 160 (warn-only rule `description-too-long`). Rewrote it to 156 chars while keeping Cass's threat-model irony intact (whitelist holds; the CI blacklist "has a bad night eventually"). Closes #681 Closes #686 Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Opus 4.8

interface · author + committer + co-author · 2026-10-06T21:07:04.000Z

build: add stage-editor.ps1 to assemble a self-contained editor folder

build: add stage-editor.ps1 to assemble a self-contained editor folder Builds RevokeCraft in Dist and copies the exe, OpenAL32.dll, mono/lib, resourses/ and projects/Example into dist/MyRevoke, pruning import libraries, debug info, intermediates and imgui.ini. Closes #37 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Opus 5

JW-Incorporated/swift2independent

a11y: solid accent fill for LockedPayoff's era chip (#3663) (#5203)

interface · author + co-author · 2026-10-06T19:25:56.000-07:00

a11y: solid accent fill for LockedPayoff's era chip (#3663) (#5203) The blurred era-label chip inside Decode's LockedPayoff painted the payoff era's raw accent as text over that same accent at ~15% tint — 3.93:1 on Midnights/evermore, under WCAG 1.4.3's 4.5:1 floor. The chip is colored by the payoff's era rather than the page's, so no surface-calibrated token can guarantee the pairing; this applies #3318's proven shape instead: solid accent fill + accentFgFor foreground, the same fix already merged for #3396/#3397/#3411. A source-lock regression test keeps the text-on-self-tint pairing from coming back. Ports attempt 1 (PR #3751, closed unmerged as superseded by #3704, which itself died stale) to the component's post-migration home in packages/ui. Tier-2: Austin — build runs Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Fable 5

JW-Incorporated/swift2independent

test: derive heroVideoFor's promoted set from the vault, not a hardcoded 11 (#5252)

interface · author + co-author · 2026-10-06T19:18:21.000-07:00

test: derive heroVideoFor's promoted set from the vault, not a hardcoded 11 (#5252) Photo Enrichment giving a video-carrying page its own MV still as the hero GROWS the promoted set by design, but video-affordance.test.ts asserted promoted.toHaveLength(11) — so every such pass failed a test the seed-only lane is not allowed to fix, and Vault Run 2026-10-05 had to revert 6 verified photos (b7fc243b) to ship green. Computing the expected set from the corpus through the public helpers (youtubeFrameId, primaryImageRef, isSubConfirmed) keeps the assertion meaningful (set equality, non-empty) while letting the corpus grow. Verified both ways: passes at today's 11 and, with b7fc243b applied locally, at 17. Closes #5192 Tier-2: Austin — build runs Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Fable 5

jjhoncv/vitrinaindependent

feat(#4): título SEO y sitemap (#15)

interface · author + co-author · 2026-10-06T18:56:19.000-05:00

feat(#4): título SEO y sitemap (#15) Closes #4 Escenarios en verde: - Título para buscadores: `/productos/taza-de-ceramica` → «Taza de cerámica | Vitrina» - Sitemap con los productos: `/sitemap.xml` incluye `/productos/taza-de-ceramica` Qué cambia: - `generateMetadata` en la landing: título «<nombre> | <proyecto>» y descripción del producto. - `app/sitemap.ts`: portada + una URL por producto. Base absoluta desde `URL` (Netlify), con fallback a `http://localhost:3000`. Cómo probarlo: `npm test && npm run build && npm run e2e`, o abrir `/sitemap.xml` y el `<title>` de una landing. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Sonnet 5.5

jjhoncv/vitrinaindependent

feat(#3): landing por producto y página de no encontrado (#14)

interface · author + co-author · 2026-10-06T18:51:02.000-05:00

feat(#3): landing por producto y página de no encontrado (#14) Closes #3 ## Escenarios en verde - «Abrir la landing de un producto» - «Producto que no existe» ## Qué cambia - `/productos/<slug>` con nombre, foto, descripción, precio y proveedor. - `notFound()` con página «Producto no encontrado» si el slug no existe. - La portada enlaza cada producto a su landing. ## Cómo probarlo `npm run e2e`, o `FUENTE_PRODUCTOS=fixture npm run dev` y abrir `/productos/taza-de-ceramica` y `/productos/no-existe`. Lint, typecheck, tests unitarios, build y e2e pasan. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Sonnet 5.5

jjhoncv/vitrinaindependent

feat(#2): portada con los productos de la hoja (#12)

browser · author + co-author · 2026-10-06T18:38:06.000-05:00

feat(#2): portada con los productos de la hoja (#12) Closes #2 **Escenario:** «Ver el catálogo» (`features/fase-1-catalogo-publico.feature`) en verde: la portada muestra Taza de cerámica y Libreta A5 con foto, nombre y precio. **Qué cambia** - `lib/productos.ts`: fuente de datos `obtenerProductos()`. Lee la pestaña `productos` con Sheets API + service account; con `FUENTE_PRODUCTOS=fixture` usa `mocks/fixtures/productos.json` (ADR 0001, ADR 0002). - Portada `app/page.tsx` con la lista de productos (dinámica, sin caché). - Playwright levanta el servidor con el fixture. **Cómo probarlo:** `npm run lint && npm run typecheck && npm test && npm run build && npm run e2e` (todo pasa; el escenario sale en verde). **Necesito de Jhonnatan** - Service account de Google Cloud con lectura a la hoja y el ID de la hoja. Configurar `GOOGLE_SHEET_ID`, `GOOGLE_SERVICE_ACCOUNT_EMAIL` y `GOOGLE_PRIVATE_KEY` en las variables de Netlify, y compartir la hoja con ese correo. - Ojo: sin esas variables, la portada en staging/producción falla (y con ella el smoke de inicio). La lectura real de Google no está probada; solo el parseo de filas y el fixture. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer web-flow · co-author claude[bot], Claude Sonnet 5.5

AlperenGuntekin/text-expansion-calculatorindependent

docs(readme): refresh topics + shotlingo cross-links

docs · author + committer + co-author · 2026-10-06T11:02:25.000Z

docs(readme): refresh topics + shotlingo cross-links Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

AlperenGuntekin/shotlingo-text-expansionindependent

docs(readme): refresh topics + shotlingo cross-links

docs · author + committer + co-author · 2026-10-06T11:02:24.000Z

docs(readme): refresh topics + shotlingo cross-links Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

AlperenGuntekin/shotlingo-screenshot-sizesindependent

docs(readme): refresh topics + shotlingo cross-links

docs · author + committer + co-author · 2026-10-06T11:02:23.000Z

docs(readme): refresh topics + shotlingo cross-links Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

AlperenGuntekin/shotlingo-locale-codesindependent

docs(readme): refresh topics + shotlingo cross-links

docs · author + committer + co-author · 2026-10-06T11:02:21.000Z

docs(readme): refresh topics + shotlingo cross-links Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

bamr87/it-journeyindependent

fix(deps): floor github-pages at >= 232 so Bundler cannot walk it back

interface · author + committer + co-author · 2026-10-06T08:20:40.000Z

fix(deps): floor github-pages at >= 232 so Bundler cannot walk it back Verified in ruby:3.4 (the Dockerfile image): with no lockfile and an unconstrained `gem 'github-pages'`, a fresh `bundle install` still resolves github-pages 222 / jekyll 3.9.0 / liquid 4.0.3 and `jekyll build` exits 1. With the floor it resolves github-pages 232 / jekyll 3.10.0 / liquid 4.0.4 and the build passes. A floor is allowed by fleet policy (no pins, no ceilings). Refs bamr87/bamr87#333 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Opus 5.5

bamr87/it-journeyindependent

fix(deps): stop committing Gemfile.lock (fleet never-commit policy)

workflow · author + committer + co-author · 2026-10-06T08:16:41.000Z

fix(deps): stop committing Gemfile.lock (fleet never-commit policy) Gemfile.lock was tracked despite .gitignore:16. Dependabot's lockfile-only bundler PR #751 satisfied a ruby-group bump by downgrading github-pages 232 -> 222 (liquid 4.0.3, which calls String#tainted?, removed in Ruby 3.2), so ci, Jekyll Build Test and dependabot-auto-merge went red every week. - git rm --cached Gemfile.lock (the .gitignore rule stays) - docker-compose.yml comments: the lockfile is local-only, not committed - AGENTS.md / copilot-instructions.md: drop the stale "must be committed" guidance in favour of the fleet policy Refs bamr87/bamr87#333 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Opus 5.5

Reservaste/frontendindependent

feat: contacto (email/teléfono) del cliente en la ficha del admin (ADR-0050, Issue #5)

implementation · author + committer + co-author · 2026-10-05T14:39:47.000Z

feat: contacto (email/teléfono) del cliente en la ficha del admin (ADR-0050, Issue #5) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

Reservaste/backendindependent

feat: contacto (email/teléfono) del cliente en la ficha del admin (ADR-0050, Issue #5)

implementation · author + committer + co-author · 2026-10-05T14:39:45.000Z

feat: contacto (email/teléfono) del cliente en la ficha del admin (ADR-0050, Issue #5) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

Reservaste/frontendindependent

Mostrar pedidos de cambio de plan en Pagos y alertar si falla la lectura (Issue #6)

implementation · author + committer + co-author · 2026-10-05T14:27:42.000Z

Mostrar pedidos de cambio de plan en Pagos y alertar si falla la lectura (Issue #6) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5.5

browser · author + committer + co-author · 2026-10-04T23:22:27.000Z

Separate user projects from the editor's own files

Separate user projects from the editor's own files The editor kept the user's content in its own install folder, so an installed editor could not save scenes, and nothing marked which files belong to a game. Adds Revoke::Project (a <Name>.mrproject YAML file plus its assets folder) with New Project / Open Project in the editor. The content browser, scene save/load and drag-and-drop now point at the open project. The renderer shader moves to resourses/shaders, and the old assets folder becomes projects/Example. Closes #34 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Opus 5

ag-grid/ag-chartsindependent

Refresh yarn.lock after the automated version bump

infra · author + committer · 2026-10-04T20:55:13.000Z

Refresh yarn.lock after the automated version bump The 14.2.0-beta.20261004 bump updated every workspace package.json but not yarn.lock. Since the Yarn 4 migration the lockfile records workspace versions, so CI's yarn install --immutable rejected the commit. Regenerate the lockfile and have bump-versions.sh refresh it on every bump.

interface · author + committer + co-author · 2026-10-04T20:38:02.000Z

Extract UUID generation into a seedable UuidGenerator class

Extract UUID generation into a seedable UuidGenerator class UUIDs were generated from module-level static RNG state in UniversallyUniqueIdentifiers.cpp, so there was no way to make generation deterministic in a test without reaching into engine internals. UuidGenerator now owns its own std::mt19937_64 and distribution. It can be constructed with an explicit seed for reproducible sequences, or default constructed to seed from std::random_device as before. UUID's default constructor wraps a single shared generator instance, so every existing call site keeps working unchanged. Also fixes UUID::Get() returning int64_t for a uint64_t member. Closes #14 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Opus 5

testing · author + committer + co-author · 2026-10-04T20:15:26.000Z

Add unit tests for the EventSystem module

Add unit tests for the EventSystem module Covers event type/name/category reporting for every event class, the data each event carries through its constructor and getters, and EventDispatcher type matching plus the Handled flag. Closes #13 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Opus 5

docs · author + committer · 2026-10-03T17:11:26.000Z

docs: accept derived artifacts on dev for v1.79.0-beta.7

docs: accept derived artifacts on dev for v1.79.0-beta.7 Производные артефакты беты — одним коммитом на dev (PROCESS.md §8, §11.4, #697). Прогон: https://github.com/Matysh/houseplan-card/actions/runs/37139464966 Скриншоты документации: отпечаток исходников; изменённые кадры: нет. Issue: #697 User-Visible: no

docs · author + committer · 2026-10-03T11:22:39.000Z

docs: accept derived artifacts on dev for v1.79.0-beta.5

docs: accept derived artifacts on dev for v1.79.0-beta.5 Производные артефакты беты — одним коммитом на dev (PROCESS.md §8, §11.4, #697). Прогон: https://github.com/Matysh/houseplan-card/actions/runs/37119340533 Скриншоты документации: отпечаток исходников; изменённые кадры: нет. Issue: #697 User-Visible: no

docs · committer · 2026-10-02T19:57:08.000+03:00

docs(reviews): manual code review of the pending-round witness (#781)

docs(reviews): manual code review of the pending-round witness (#781) Manual review requested by the owner after the automated reviewer failed. Material e931a949: green, no High/Medium findings, one non-blocking Low. Issue: #781 User-Visible: no

author Codex · committer claude[bot]

testing · committer · 2026-10-02T19:57:08.000+03:00

test(process): поймать оживление прошлого запроса (#781)

test(process): поймать оживление прошлого запроса (#781) Issue: #781 User-Visible: no

reecejunior/techzimintindependent

Log 2026-10-02 growth/product strategy notes

implementation · author + committer + co-author · 2026-10-02T11:47:51.000Z

Log 2026-10-02 growth/product strategy notes Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5

reecejunior/techzimintindependent

Fix raw ISO timestamp shown on reviewer profile review dates

context · author + committer + co-author · 2026-10-02T11:47:18.000Z

Fix raw ISO timestamp shown on reviewer profile review dates ProfileClient's review history rendered r.createdAt verbatim, e.g. "2024-06-01T12:34:56.789Z", instead of a readable date. The identical Review.createdAt field is already formatted with .slice(0, 10) in StartupDetailClient's ReviewCard — this applies the same fix here so profile pages are consistent with startup detail pages. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

author claude[bot] · committer claude[bot] · co-author Claude Sonnet 5

docs · author + committer · 2026-10-02T10:47:04.000+03:00

docs: review document for #780

docs: review document for #780 Issue: #780 User-Visible: no

interface · author + committer · 2026-10-02T10:18:38.000+03:00

docs(reviews): rebuild the review index after rebase onto dev

docs(reviews): rebuild the review index after rebase onto dev

workflow · author + committer · 2026-10-02T10:18:29.000+03:00

fix(process): prepare and the usage step run pipeline scripts from one pinned dev snapshot (#765)

fix(process): prepare and the usage step run pipeline scripts from one pinned dev snapshot (#765) The body of _process.yml is read from dev (@dev, #623). After "Перейти на ветку задачи" the working copy of job prepare is the task branch, and a show/ship branch with a clean merge is not rebased before review: its scripts/ may lag dev or be replaced. #749 fixed job integrate; prepare still ran four control scripts from the material — the issue-body digest for the anchor, --reuse of a green verdict (#499), validate-gate (#510) and the spec-change check (#517). The material decided its own admission: a branch whose review-doc-guard.mjs prints reuse=true merges without the model. model_review took model-usage.mjs from the material too: a lagging branch has none, and the publication silently wrote reason=missing. Now prepare extracts one snapshot right after setup-node, before the branch switch: `git rev-parse origin/dev` once and `git archive <sha> scripts .github/workflows/validate.yml`, so the git fetch of the track and rebase steps cannot mix versions. Every repo script of the job runs from it via TOOLS — the track step and the rebase guard lose their own extractions. The SHA goes out as job output tools_sha; the usage step of model_review archives the same commit inside itself, so a snapshot failure is a failure of the reporting step (continue-on-error), not of the stage. The model session runs on that runner, so the usage line stays untrusted input parsed strictly (#556, #737). withMaterialAnchors is idempotent: a repeated call drops the separator the previous call wrote instead of piling up `---` lines. Tests: test/process-prepare-tools.test.mjs — the job contract (no step calls scripts/ from the working copy, one pinned archive before the branch switch, tools_sha reaches model_review) and the steps as they are, on real bash and git: a branch behind dev without model-usage.mjs and with a substituted review-doc-guard.mjs; the anchor digest, reuse and the spec check come from dev, the usage line is data even after dev moved. Red on the old workflow: all four. Harnesses of process-track, rebase-generated, review-doc-guard and model-usage take the prepare snapshot. Three registry mutants (reuse from the material, usage from moving dev, separators). Canon: PROCESS.md §10.4 «Скрипты конвейера — из dev» covers prepare and the usage step; «Расход модели» names it a pipeline step, not the reviewer's. Issue: #765 User-Visible: no

docs · committer · 2026-10-02T10:18:29.000+03:00

docs(reviews): independent code review of pipeline snapshot (#765)

docs(reviews): independent code review of pipeline snapshot (#765) Manual review requested by the owner after the automated reviewer failed. Material af09d36d: green, no High/Medium findings, one non-blocking Low. Issue: #765 User-Visible: no

docs · author + committer · 2026-10-02T10:09:11.000+03:00

docs: review document for #780

docs: review document for #780 Issue: #780 User-Visible: no

docs · author + committer · 2026-10-01T23:05:23.000Z

docs(reviews): индекс после сдвига каталога (#744)

docs(reviews): индекс после сдвига каталога (#744) Issue: #744 User-Visible: no

docs · author + committer · 2026-10-01T23:05:20.000Z

docs: review document for #744

docs: review document for #744 Issue: #744 User-Visible: no

implementation · committer · 2026-10-01T21:00:48.000Z

fix: keep warm editor adoption subordinate to user navigation (#762)

fix: keep warm editor adoption subordinate to user navigation (#762) Issue: #762 User-Visible: yes

testing · committer · 2026-10-01T21:00:48.000Z

test: await warm navigation request before checking its authority (#762)

test: await warm navigation request before checking its authority (#762) Issue: #762 User-Visible: no

docs · author + committer · 2026-10-01T21:00:48.000Z

docs(reviews): индекс после сдвига каталога (#762)

docs(reviews): индекс после сдвига каталога (#762) Issue: #762 User-Visible: no

docs · author + committer · 2026-10-01T21:00:48.000Z

docs: review document for #762

docs: review document for #762 Issue: #762 User-Visible: no

implementation · committer · 2026-10-01T21:00:48.000Z

refactor: keep room warm-draft construction in the lazy editor (#762)

refactor: keep room warm-draft construction in the lazy editor (#762) Issue: #762 User-Visible: no

docs · author + committer · 2026-10-01T20:52:49.000Z

docs: review document for #662

docs: review document for #662 Issue: #662 User-Visible: no

docs · author + committer · 2026-10-01T20:33:15.000Z

docs: review document for #662

docs: review document for #662 Issue: #662 User-Visible: no

docs · author + committer · 2026-10-01T18:38:03.000Z

docs(reviews): индекс после сдвига каталога (#740)

docs(reviews): индекс после сдвига каталога (#740) Issue: #740 User-Visible: no

docs · author + committer · 2026-10-01T18:38:00.000Z

docs: review document for #740

docs: review document for #740 Issue: #740 User-Visible: no

browser · author + committer + co-author · 2026-10-01T18:27:53.000Z

perf(stairs): draw all treads of a stair as one path (#740)

perf(stairs): draw all treads of a stair as one path (#740) A floor with stairs pays for them on every switch to it: the stair layer is emptied on other floors, so Lit recreates every symbol on each return and the browser lays out and paints it again. With 250 stairs (the large-house fixture, the per-floor limit) that was 2,875 SVG elements and about 40 ms per entry locally; each stair carried 3-7 separate tread lines with four bound coordinates each. The treads of one stair are now a single <path class="hp-stair-tread"> with one `M a L b` subpath per tread, in geometry order and with the numbers the lines carried. Treads of one stair never overlap (straight: parallel, >= 20 cm apart; spiral: inner ends >= 6.7 cm apart at the 3.6 cm stroke), so the path paints the same pixels at any opacity. The outline points and the tread data are built once per cached geometry object (cachedStairMarkup, weak keys), not on every render. The View and plan-editor layers share the strings; outline, hit polygon, trapezoid, arrow, attributes and handlers are unchanged. Floor 1 of the fixture drops from 4,210 to 2,960 elements. Witnesses: the unit test for the path data and its cache, and the smoke_stairs markup checks in View and in the plan editor, are red on dev. The stairs-view-tread-lines mutant restores the View lines. Issue: #740 User-Visible: no Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd

author claude · committer claude[bot] · co-author Claude Opus 5.5

browser · author + committer + co-author · 2026-10-01T18:27:53.000Z

docs(screenshots): source fingerprint for #740 on the current dev — all frames pixel-identical

docs(screenshots): source fingerprint for #740 on the current dev — all frames pixel-identical Validate on the conveyor's rebase eb439e81 failed preflight: the branch changes visual sources (stairs), so the screenshot check is strict on it, and the fingerprint was stale against the moved dev (#739, #742, #759). The branch is rebased onto b84465e5 (inventory counts merged with #742: lifecycle 89, total 204/200) and `docs:accept --identical` re-captured all 11 frames: pixel-identical, only the source fingerprint moves. Issue: #740 User-Visible: no Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd

author claude · committer claude[bot] · co-author Claude Opus 5.5

docs · author + committer · 2026-10-01T18:27:53.000Z

docs: review document for #740

docs: review document for #740 Issue: #740 User-Visible: no

docs · author + committer · 2026-10-01T17:53:24.000Z

docs: review document for #762

docs: review document for #762 Issue: #762 User-Visible: no

browser · author + committer + co-author · 2026-10-01T16:49:55.000Z

fix(process): reconcile canon and hints with the pipeline after #707–#730 (#748)

fix(process): reconcile canon and hints with the pipeline after #707–#730 (#748) Five places still described the pipeline as it was before code that is already in dev: - the S3 hint of the task packet told the author to push the branch, while the spec lives in the issue body (§2.3, #517) and nothing is pushed before S5 (§11.8); - process-gate printed «FAIL п.9 Gates: light» for a trailer nobody writes or reads, while §10.2 item 9 is the unimplemented release:prerelease verdict check. The check is removed; a contract test ties every RULES key to an implemented item of §10.2 and every finding number to a RULES key; - §10.4 item 4 demanded a heredoc in run:, while #723/#730 and their tests demand the opposite: commit messages echo line by line into a file, comment and summary texts come from code; - the ship merge comment, AUTHOR.md, REVIEWER.md and AGENTS.md named only the pre-beta document, though since #727 the night reads ship code first; - the nightly publication committed «docs: ship review for nightly … перед бетой» with the beta step's Issue: #696. It now has its own subject (the document name), body and Issue: #727; the beta message is unchanged. The browser-guard inventory note still said growth above 200 fails mutation-gate --check; since #699 it is a guideline and --check warns. Its counts now match the inventory: 205, lifecycle 90. Issue: #748 User-Visible: no Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd

author claude · committer claude[bot] · co-author Claude Opus 5.5

context · author + committer + co-author · 2026-10-01T16:49:55.000Z

docs(reviews): индекс после сдвига каталога (#748)

docs(reviews): индекс после сдвига каталога (#748) Issue: #748 User-Visible: no Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018qZfe7YS4rqEMKoVeS3GKd

author claude · committer claude[bot] · co-author Claude Opus 5.5