Usage Pace 0.4.5 — compact quota details

Patch release for a quieter usage dialog and clearer reset status.

Changes

• Quota details first. Removed the token-total card and its explanatory text. The UI no longer fetches or polls token totals; bb usage-pace --tokens and the token RPC remain available.
• Only relevant accounts. The strip, dialog, pools, and reset-read targets share an account filter. Uninstalled providers and discoveries without quota windows, account identity, or a previous reading stay hidden. This removes unused Cursor/OpenCode entries without hiding real accounts whose reads fail or logins expire. Known accounts without quota windows keep their reset details and never claim their quota will last.
• Compact header. Title, update time, and refresh share one row. Mobile safe-area padding and accessible dialog behavior are preserved.
• Clearer reset status. Conditional five-hour availability shows the provider-reported reason directly, such as “Not available · session limit not reached.” Removed duplicate zero-balance copy, repeated disclaimers, and nested reset cards. Saved balances, expirations, unknown states, and conditional availability remain distinct.

This does not change provider eligibility or redeem any resets. The read-only live check for the reported Claude account returned zero saved resets and not_at_wall; no missing grant was reproduced. Raw CLI account output remains available for diagnosis.

Minimum requirements and dependency pins are unchanged: BB 0.44.0 / SDK 0.5.29.

Verification

• 425 Vitest tests and 3 token-total tests pass, including before/after regressions for unused providers, known-account failures, token polling, and reset explanations.
• Type checking, comment-policy checks (114 files, zero violations, all 13 regression tests), staged policy, the ordinary pre-commit hook, redacted secret scanning, and production-only rebuild/runtime smoke checks pass. Release artifacts use BB 0.44.0 / SDK 0.5.29; the isolated install uses BB's bundled npm 11.16.0.
• Synthetic browser checks at 1200px, 375px, and 320px confirm no horizontal overflow or token requests. Header height falls from 40.75px to 24.5px in the component fixture. Refresh, Escape/reopen, empty/stale accounts, and available/unknown reset states pass without browser errors.
• The browser fixture uses production components and generated CSS with synthetic accounts/RPC and placeholder provider marks. It is not a live installation or native desktop acceptance test.

Install or upgrade

For a new installation:

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.5'
bb plugin config usage-pace set showStrip true

Reload the app UI to show the optional footer card. Use @^0.4.5 instead of the exact tag to track compatible releases.

For an existing installation, inspect its source first:

bb plugin source usage-pace

Branch-tracking and compatible release-range installs can update while preserving settings:

bb plugin update usage-pace

Exact tags and commits remain pinned. Changing a pinned source may require removal and reinstall; removal deletes settings, secrets, and schedules, so record the configuration first. See upgrade instructions.

PR CI, resulting-main CI, and tag CI passed. The annotated v0.4.5 tag points to merge commit a1d2c89b4c60b5f60b5342e68adf1461e1eb057b, whose tree exactly matches tested PR head c661bdedda287a9456e57ff446f9dd35508e6702. GitHub Actions lists no runs for those SHAs; hosted CI was verified through Woodpecker commit statuses.

Publishing this release does not install, update, reload, or reconfigure any live plugin.

PR #11 · Full changelog

Usage Pace 0.4.4 — macOS sidebar footer fix

Patch release for the compact strip in bb 0.47's macOS desktop app.

Fixes

• The strip uses bb's supported sidebar footer API. Its own footer icon opens a bordered, scrollable card above bb's footer buttons. The strip no longer inserts into bb's footer markup, where the desktop app's fixed 48px row hid account rows and pushed controls below the window.
• Opens once, not after every refresh. With showStrip enabled, the card opens at startup. bb allows one footer disclosure open at a time: opening Provider usage closes the strip; the Usage Pace icon brings it back. Background updates never take over the open card. bb does not remember the open disclosure across reloads.
• Refresh and reset reads follow the right lifetimes. Quota polling every five minutes while the page is visible, focus/visibility refreshes, and refresh after an exhausted account's reset passes continue with the strip closed. Saved-reset reads run only while the strip or details dialog is open, sharing the same inventory and badges.

showStrip remains off by default. Reload the app UI after changing it so bb can add or remove the registered footer item. The dialog, Provider usage pace ticks/per-tab errors/hidden providers, server-only filtering, CLI, and read-only reset behavior are unchanged. Release artifacts retain the BB 0.44.0 / SDK 0.5.29 toolchain; minimum requirements and dependency pins are unchanged.

Verification

• The same startup regression failed on pristine 0.4.3 and passed after the fix: one supported disclosure, opened once, with no injection into bb's footer. All five new lifecycle tests pass, including off/no registration, closed-card quota refresh, passed-reset refresh, and saved-reset reads only while the actual strip or dialog is open.
• All 411 Vitest tests, 3 token-total tests, TypeScript, and release validation passed and were rerun before merging. Comment policy passed across 113 files with zero violations; all 13 policy regression tests passed. Staged policy, the ordinary pre-commit hook, and redacted secret scanning passed during preparation.
• Production-only rebuilds and release validation passed with BB 0.44.0 / SDK 0.5.29 and BB's bundled npm 11.16.0, including version/SDK metadata, host digest, and isolated backend/host smoke loading. No provider requests were made by release validation.
• A throwaway bb-app 0.47.1 instance with synthetic example.test usage reproduced the macOS layout failure on 0.4.3. The actual release bundle passed at 1440×900 and 1440×600, in both stubbed macOS desktop mode and normal browser tabs: all five strip rows and all footer controls are inside the window and centre-hit-testable. Opening Provider usage closes the strip; the Usage Pace icon reopens it. The bridge stub leaves missing numeric inset fields undefined so bb uses its shipped defaults. No bb code or layout CSS is patched. The temporary instance and its data were removed afterwards.
• The existing component/CSS fixture also passed server/all-host strip widths 172/212/252px and desktop/mobile dialogs 1200/375px, preserving stale-account explanations, partial pools, saved-reset badges, and server-only reset targeting without horizontal overflow.
• PR CI, resulting-main CI, and release-tag CI passed. The annotated v0.4.4 tag points to merge commit 842ae6175dbe7f6a7773b1b2f892a881fbfebad5, whose tree exactly matches reviewed and tested PR head 3a5cec7891c3a98548cd147136152a6bc74da16e. GitHub Actions lists no runs for those SHAs; Woodpecker results were verified through commit statuses.
• Not yet verified: installation or upgrade of 0.4.4 on either real bb server, native AppKit behavior, or real provider acceptance. Before/after measurements and screenshots are recorded in sidebar footer verification. Only synthetic identities and example.test addresses are used.

New installations

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.4'
bb plugin config usage-pace set showStrip true

Reload the app UI to show the optional card. Use @^0.4.4 instead of the exact tag to track compatible releases.

Upgrades

Inspect the installed source first:

bb plugin source usage-pace

For a branch-tracking install whose branch contains 0.4.4, or a release-range install admitting this tag, update in place and preserve settings:

bb plugin update usage-pace

Exact tags and commits do not advance to a newer tag. Moving a pinned install, or changing an existing install to a tag or range, requires removal first. Removal deletes settings, secrets, and schedules. Record your configuration and restore it after reinstalling:

bb plugin config usage-pace
bb plugin remove usage-pace
bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.4'

Use @^0.4.4 in the last command for compatible-release tracking. Do not install over an existing managed install. Reload the app UI after upgrading.

Publishing this release does not install, update, reload, or reconfigure plugins on either live bb server. The user's old checkout and other repositories remain untouched.

PR #10 · Full changelog

Usage Pace 0.4.3 — honest stale readings

Patch release for usage sources that keep returning their last successful reading when they cannot refresh an account.

Fixes

• Old readings no longer look current. Normal polling still accepts cached observations. Once a reading reaches the configured usage-cache period (five minutes by default), Usage Pace checks that resource with refresh: true, at most once per cache period. If the forced response is still at least one cache period old at request start, or has no observation, the account is stale—even before a reset passes. Recent observations within the period are accepted without requiring an exact request-time timestamp. Manual refresh bypasses the probe limit.
• Stale quota explains what happened. The strip tooltip, dialog, and CLI show the last successful reading's age, name the source that couldn't refresh the account, and include its reason when supplied. They no longer recommend another refresh immediately after that failure. Stale or unobserved quota remains “usage unknown,” without numeric pace, a zero-usage implication, or a stale forecast. Historical percentages remain in JSON with stale: true and null pace.
• Known accounts keep their history through authentication failures. expired and unauthenticated results retain the account's prior windows, last-success timestamp, and omitted identity metadata, marked stale with a source-named state/reason. Retained accounts remain visible and count as unknown in their pool. A failed freshness probe also retains the cached reading it just received. Unknown identities stay separate; explicit identity changes never inherit another account's history.
• Healthy duplicates still win. The 0.4.2 merging rule is unchanged: a stale or failed source cannot replace a healthy measurement of the same issued account identity. Other hosts, sources, and failure messages remain visible.

The full freshness rule is documented in README. This does not renew credentials or fix the upstream usage source. Release artifacts remain built with BB 0.44.0 / SDK 0.5.29; minimum requirements and dependency pins are unchanged. Saved resets remain read-only; the CLI JSON accounts/providers alias, token totals, and Codex-only --resets behavior are unchanged.

Verification

• The four forced/unforced before- and after-reset regression cases failed against pristine 0.4.2 and passed with this fix. The full collector/account/integration regression set passed all 38 tests, including healthy on-demand refresh, weekly-only stale readings, probe throttling, retained source reasons/history, and healthy duplicates winning.
• All 406 Vitest tests, 3 token-total tests, TypeScript, and release validation passed. TypeScript, the complete test suite, and release validation were rerun before merging.
• Comment policy passed across 98 files with zero violations; all 13 policy regression tests passed. Staged policy, the ordinary pre-commit hook, and redacted secret scanning passed during preparation.
• Production-only rebuilds and release validation passed with BB 0.44.0 / SDK 0.5.29 and BB's bundled npm 11.16.0, including version/SDK metadata, host digest, and isolated backend/host smoke loading. No provider requests were made by release validation.
• Synthetic desktop/mobile browser checks showed B's last-read age, named source failure, and unknown quota without numeric pace or retry advice, alongside healthy A/Codex, the partial Claude pool, correct locations, and matching read-only resets. Strip widths 172/212/252px and desktop/mobile dialogs 1200/375px passed without horizontal overflow.
• PR, resulting-main, and release-tag Woodpecker CI passed. The annotated v0.4.3 tag points to merge commit abe551ced35a827d2d92041b2b798ec20790dc96, whose tree exactly matches the reviewed and tested PR head.
• Not yet verified: live installation or upgrade of 0.4.3 on either bb server. Only synthetic identities and example.test addresses are used in the regression evidence, recorded in multi-account verification.

New installations

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.3'
bb plugin config usage-pace set showStrip true

The second command enables the optional strip. Use @^0.4.3 instead of the exact tag to track compatible releases.

Upgrades

Inspect the requested and resolved source first:

bb plugin source usage-pace

For repository/branch-tracking installs whose branch contains 0.4.3, or a release-range install that admits this tag, update in place and keep settings:

bb plugin update usage-pace

An exact tag or commit does not advance to a newer tag with update. Moving a pinned install, or changing an existing install to a tag or range, requires removal first. Removal deletes settings, secrets, and schedules. Record the current configuration and restore your settings after reinstalling:

bb plugin config usage-pace
bb plugin remove usage-pace
bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.3'

Use @^0.4.3 in the final command for compatible-release tracking. Do not use plugin install as an in-place upgrade: bb 0.46.1 rejects the duplicate plugin ID.

Publishing this release does not install, update, reload, or reconfigure plugins on either live bb server.

PR #9 · Full changelog

Usage Pace 0.4.2 — account-merging fixes

Patch release for the remaining 0.4.1 account-merging gap with bb 0.46.1's cached last-success timestamps on failed reads.

Fixes

• Healthy quota wins over failed duplicates. Merged accounts prefer the newest healthy (status: "ok", not stale) measurement, even when a failed source reports a newer last-success timestamp. A transient machine read failure no longer makes a healthy merged account stale or turns a two-account pool into 1/2 available · 1 unknown.
• Known accounts keep their metadata through errors. When an error keeps the resource's known identity but omits email, plan, or plan label, retain the cached fields. Email-based aliases and plan weights remain available. Explicit identity changes still discard the old account's metadata and windows; never-known identities remain separate.
• Failures remain visible and honest. Merged hosts, sources, and source-error messages remain visible. An account whose only source fails stays stale and counts as unknown despite retaining its alias and plan. When all sources are stale, the newest observation still wins.

Release artifacts remain built with BB 0.44.0 / SDK 0.5.29; minimum requirements are unchanged. Saved resets remain read-only. The CLI JSON accounts/providers alias, nullable pace, token totals, and Codex-only --resets behavior are unchanged.

Verification

• The final collector/merging test set produced 7 failures with 0.4.1's source and passed all 22 tests with this fix. Coverage includes realistic last-success timestamps, retained aliases/plan metadata, sole-source failures remaining unknown, unchanged all-stale ordering, and explicit login changes.
• All 390 Vitest tests, 3 token-total tests, and TypeScript passed. TypeScript, the complete test suite, and release validation were rerun before merging.
• Comment policy passed across 96 files with zero violations; all 13 policy regression tests passed. Staged policy, the ordinary pre-commit hook, and redacted secret scanning passed.
• Production-only rebuilds and release validation passed with BB 0.44.0 / SDK 0.5.29 and BB's bundled npm 11.16.0, including version/SDK metadata, host digest, and isolated backend/host smoke loading. No provider requests were made by release validation.
• Synthetic desktop/mobile browser checks retained the 2/2 pool, alias, source-error message, matching saved-reset badge, view-specific locations, and reset-reader provenance without compact-strip/dialog overflow.
• PR, resulting-main, and release-tag Woodpecker CI passed. The annotated release tag points to merge commit d7bdcf5b16f675491726c59b6bb96cf767041de9, whose tree exactly matches the reviewed and tested PR head.
• Not yet verified: live installation or upgrade of 0.4.2 on either bb server. The reproduction and before/after evidence use only synthetic identities and example.test addresses and are recorded in multi-account verification.

New installations

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.2'
bb plugin config usage-pace set showStrip true

The second command enables the optional strip. Use @^0.4.2 instead of the exact tag to track compatible releases.

Upgrades

Inspect the requested and resolved source first:

bb plugin source usage-pace

For repository/branch-tracking installs whose branch contains 0.4.2, or a release-range install that admits this tag, update in place and keep settings:

bb plugin update usage-pace

An exact tag or commit does not advance to a newer tag with update. Moving a pinned install, or changing an existing install to a tag or range, requires removal first. Removal deletes settings, secrets, and schedules. Record the current configuration and restore your settings after reinstalling:

bb plugin config usage-pace
bb plugin remove usage-pace
bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.2'

Use @^0.4.2 in the final command for compatible-release tracking. Do not use plugin install as an in-place upgrade: bb 0.46.1 rejects the duplicate plugin ID.

Publishing this release does not install, update, reload, or reconfigure plugins on either live bb server.

PR #8 · Full changelog

Usage Pace 0.4.1 — live-check fixes

Patch release fixing all seven issues found during the 0.4.0 live check on bb 0.46.1 / SDK 0.6.49 with two Claude Accounts logins.

Fixes

1. Healthy server accounts stay visible. Account usage errors, rejected resource reads, and failed source inventories no longer mark an entire host unavailable. Server-only filtering only treats host identity, availability, and host-list failures as server failures.
2. Account errors stay on their rows. The dialog footer and CLI warning no longer repeat individual account errors. Discovery, host-list, and source-list failures still produce collection warnings.
3. Card errors stay on the failed tab. Recognizes all three bb 0.46.1 notices, including “Couldn’t refresh. Showing last update.”, while retaining older wording. Healthy Codex, Claude Code, and All accounts tabs stay clear when the failed provider is known.
4. Merged accounts show the right location. Strip descriptions and dialog locations use the server machine in server-only mode and all associated machines otherwise. Saved-reset details name the machine that actually supplied the verified inventory.
5. No repeated provider labels. CLI, strip, and dialog omit fallback account labels that equal the provider display name, ignoring case. Meaningful account labels and aliases remain.
6. Separate installation and upgrade instructions. Inspect the source before choosing an in-place update or a settings-destructive remove/reinstall.
7. Transient keyless errors do not split known accounts. A resource that previously reported an account key keeps it when an error omits the key, so a sleeping laptop does not add a phantom pool member. Never-known identities remain separate; a successful login change replaces the key.

Minimum requirements remain BB 0.44.0 / SDK 0.5.29. Committed release artifacts were built with that toolchain. Saved resets remain read-only; CLI account/provider JSON aliases, nullable pace, token totals, and Codex-only --resets behavior are unchanged.

Verification

• 385 Vitest tests, 3 token-total tests, and TypeScript passed.
• Comment policy passed across 95 files with zero violations; all 13 policy regression tests passed. Staged secret scanning passed before the implementation commit.
• Production-only rebuilds and release checks passed, including BB's bundled npm 11.16.0, version/SDK metadata, host digest, and isolated backend/host smoke loading. No provider requests were made by release validation.
• Synthetic desktop/mobile browser checks covered current-view account labels, reset-reader provenance, nonduplicate headings, and compact/dialog overflow.
• PR, resulting-main, and release-tag Woodpecker CI passed. The release tag points to the merged main commit containing the tested tree.
• Not yet verified: live installation or upgrade of 0.4.1 on either bb server. The completed 0.4.0 live observations and synthetic regression evidence are recorded in multi-account verification.

New installations

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.1'
bb plugin config usage-pace set showStrip true

The second command enables the optional strip. Use @^0.4.1 instead of the exact tag to track compatible releases.

Upgrades

Inspect the requested and resolved source first:

bb plugin source usage-pace

For repository/branch-tracking installs whose branch contains 0.4.1, or a release-range install that admits this tag, update in place and keep settings:

bb plugin update usage-pace

An exact tag or commit does not advance to a newer tag with update. Moving a pinned install, or changing an existing install to a tag or range, requires removal first. Removal deletes the plugin's settings, secrets, and schedules. Record the current configuration and restore your desired settings after reinstalling:

bb plugin config usage-pace
bb plugin remove usage-pace
bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.1'

Use @^0.4.1 in the final command for compatible-release tracking. Do not use plugin install as an in-place upgrade: bb 0.46.1 rejects the duplicate plugin ID.

Publishing this release does not install, update, reload, or reconfigure plugins on either live bb server.

PR #6 · Full changelog

Usage Pace 0.4.0 — multiple accounts and provider pools

Multiple quota accounts

• Discover every provider-usage.v1 source, including machine-local providers, Account Pooler, Claude Accounts, and Usage Pace's Grok source.
• Merge accounts only by provider-issued identity, keeping the freshest measurement and all reporting machines/sources. Matching emails do not merge accounts; unknown identities remain separate.
• Show one account row per quota identity plus a provider pool when at least two accounts are displayed. Add account aliases, plan-weight overrides, availability, observation age, and separate model-family exhaustion.
• Keep shared-source accounts visible in the server-only view. Failed sources retain marked-stale measurements; stale or unobserved usage is not shown as zero.

Pool weekly use and pace are plan-weighted estimates, marked partial when measurements or weights are missing. Weekly quotas may not scale exactly with session plan multipliers. There is no pool run-out forecast.

Fixes and CLI

• Restore card annotations using the canonical Provider Usage RPC route and plural providerIds. Match annotations by machine, provider, and account section; ambiguous rows draw nothing.
• Match saved-reset inventory to the quota account using Claude's profile UUID or Codex's account ID, never email. Unread or mismatched inventories have no badge; a verified failed read remains unknown. No resets are redeemed and no credentials are changed.
• bb usage-pace prints pools before accounts. JSON adds accounts[] and pools[]; providers[] remains a deprecated account alias through 0.4.x. Each window retains nullable pace.

Verification and limitations

• 363 Vitest tests, 3 token-total tests, TypeScript, zero comment-policy violations, and all 13 policy regression tests passed.
• BB 0.44.0 / SDK 0.5.29 prebuilt artifacts, production-only source rebuild, release metadata, host digest, and isolated bundle loading passed. Compatibility floors are unchanged.
• Read-only live collection merged one Codex identity across three machines. Browser fixtures covered three Claude accounts, a pool, account-safe reset badges, compact widths, and mobile details.
• Still unverified: a live bb 0.46.1 installation with Claude Accounts serving at least two logins. Neither shared account source was installed in the development environment; shared-source fixtures are representative contracts, not live captures. See verification.

Install or upgrade

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.4.0'

Choose @^0.4.0 to track compatible 0.4 releases with bb plugin update usage-pace. Exact older tags and ^0.3.x ranges do not automatically advance to this new minor version; select the new tag or range explicitly. Publishing this release does not install or reload the local plugin.

Full changelog

Usage Pace 0.3.10 — confirmed zero Claude resets

Fix

Claude's saved-reset badge now shows ↺ 0 instead of ↺ ? when the provider explicitly reports no_grant with an empty grant list. Previously, the adapter required eligible: true and incorrectly treated this confirmed absence as an unreadable inventory.

Missing, malformed, contradictory, or access-gated inventory still shows unknown rather than a guessed zero. Conditional five-hour reset availability remains separate from the saved balance. No resets are redeemed and no authentication is changed.

Verification

• 330 Vitest tests, 3 token-total tests, 13 comment-policy regression tests, TypeScript, and comment-free source checks passed.
• The original bug was reproduced against the authorized Claude account over SSH. Both the fixed reader and regenerated host bundle returned a confirmed zero using read-only GET requests.
• Production-only source rebuild, release metadata, host digest, and isolated backend/host loading passed with BB's bundled npm.
• Prebuilt artifacts preserve the BB 0.44.0 / SDK 0.5.29 baseline. Compatibility requirements and existing settings are unchanged.

Install

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.3.10'

Compatible-range installations can update with bb plugin update usage-pace. Installations pinned to an exact older tag need to select the new tag.

Full changelog

Usage Pace 0.3.9 — clearer quota forecasts

Changes

• Separate current quota from forecasts. near limit now means at least 80% of a quota is already used. A forecast-only warning says high projected use, so 5% used early in the week is no longer mislabeled as nearly exhausted.
• Show the projected shortfall before reset. The compact warning now says 4h short instead of a distant countdown such as out in 6d 7h. Hover for the quota window and run-out time.
• Keep pace and shortfall anchored to the last usage report, while tooltip countdowns continue ticking. Hidden model limits, stale-data handling, estimated-now warnings and confirmed-exhaustion reset countdowns are preserved.

No changes to quota estimation, risk thresholds, saved resets, configuration or compatibility requirements.

Verification

• 323 application tests, 13 comment-policy regression tests, TypeScript and comment-free source checks passed.
• Verified the installed sidebar in local BB with real provider data; the preview was user-approved.
• Production-only source builds, release metadata, host digest and isolated backend/host loading passed using BB's bundled npm and SDK runtime.
• Prebuilt artifacts preserve the BB 0.44.0 / SDK 0.5.29 baseline. Package contents exclude prototype artifacts.

Install

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.3.9'

Compatible-range installations can update with bb plugin update usage-pace. Existing settings are preserved.

Full changelog

Usage Pace 0.3.8 — compact reset badge alignment

Fix

Align the compact strip's saved-reset badge with the provider icon and first quota readout. Previously, different button heights put the reset balance 3.5px too high.

The badge remains anchored to the first line for weekly/session rows and rows with extra warnings. Other reset-badge surfaces retain their existing sizing. No changes to quota calculation, saved-reset inventory, configuration or compatibility requirements.

Verification

• 305 application tests, 13 comment-policy regression tests, TypeScript and comment-free source checks passed.
• Browser geometry checks passed for 42 desktop/mobile combinations at actual content widths of 172, 212 and 252px, with zero vertical offset and no visible horizontal overflow.
• Production-only source builds, release metadata, host digest and isolated backend/host loading passed using BB's bundled npm and SDK runtime.
• Prebuilt artifacts preserve the BB 0.44.0 / SDK 0.5.29 baseline. Package contents exclude prototype artifacts.

Install

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.3.8'

Compatible-range installations can update with bb plugin update usage-pace. Existing settings are preserved.

Full changelog

Usage Pace 0.3.7 — pace-first compact strip

Changes

• Lead the optional compact strip with pace multipliers, keeping quota used secondary: W 0.5× pace · 11% used. Weekly and session windows have separate, labelled lines; the ordinary status dot no longer competes with the primary number.
• Explain 1.0× even pace and the window-average meaning. Numeric pace stays anchored to its usage snapshot while countdowns continue ticking; it is not recent burn or an exact agent-count recommendation.
• Preserve all-window constraints, including hidden model limits. A near-limit window still warns when its reset or pace is unknown. Missing, early, stale and incomplete data cannot produce a reassuring numeric pace claim.
• Retain projected-exhaustion estimates, confirmed-exhaustion reset countdowns, refresh-once/awaiting-refresh behaviour, account grouping, keyboard access and separate saved-reset controls.
• Keep narrow sidebars compact with container-aware spacing. Preserve optional-strip visibility, server-only filtering, saved-reset inventory, collection/authentication behaviour, the CLI and the separately owned Provider Usage card.

Verification

All 305 application tests, 13 comment-policy regression tests, TypeScript checking, comment-free source checks, production-only source builds using BB's bundled npm, release metadata/digests and isolated backend/host loading passed. Package contents contain no prototype variants, switcher or static server.

Prebuilt artifacts retain the existing BB 0.44.0 / Plugin SDK 0.5.29 compatibility baseline. The new layout was verified in the actual BB sidebar at narrow desktop/mobile widths, and tried and approved in the running BB instance on devinabox.

Install

bb plugin install 'git:https://github.com/scowalt/bb-plugin-usage-pace.git@v0.3.7'
bb plugin config usage-pace set showStrip true

Existing compatible-range installations (@^0.3.6 or @^0.3.7) can update with bb plugin update usage-pace. Existing server-only and other plugin settings are unchanged. Requires BB 0.44.0+ with Plugin SDK 0.5.29+.

Full changelog

GitHub comment-free source workflow passed on release commit b99b54e.