fix(web): usage totals no longer jump while devices report in

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}

[t3dotgg]

Opening the usage page with several connected devices was gross: totals rendered as soon as the first device answered, then every number on the page revised itself as each remaining device landed.

The page now holds its content until every device is terminal (reported, failed, or stale). While waiting it keeps the header toggles live and shows a static skeleton in the page's final shape, topped with a per-device strip: reported devices get a check, failed ones an x, and still-scanning ones a duty-cycled pulse. The strip only appears with two or more devices; a single-device user just sees the skeleton. Failed and stale devices never block the gate, and changing the window or hitting refresh restarts it.

No shimmer or continuous animation; the skeleton is static and the scanning pulse reuses the existing duty-cycled status-pulse keyframes.

Built by Claude Fable 5 via Claude Code.

🤖 Generated with Claude Code

---

Note

Low Risk
UI-only loading and presentation changes on the usage page; no changes to how usage is fetched or merged.

Overview
Fixes jumping usage numbers when several connected environments report at different times. The page no longer renders merged totals while any device is still scanning; it waits until every environment is terminal (success, error, or excluded as stale).

While loading, the header stays interactive and the body shows a static skeleton shaped like the final layout (no shimmer). With two or more devices, a per-device strip shows checkmarks for finished devices, an X for failures, and a pulsed label for devices still scanning, plus a count of devices left.

Coverage messaging moves behind the gate: the “partial totals” line is removed because in-progress devices never show real numbers anymore. Failed, stale, and duplicate-transcript warnings still appear once the page settles.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix usage totals jumping on the Usage page while devices report in

• Replaces the previous partial-totals display with a static skeleton layout while usage data is still collecting from environments, preventing layout shifts as devices report in.
• Adds a UsageDeviceStrip component (shown when multiple environments are present) that displays per-device status (completed, failed, or scanning) and a count of devices still scanning.
• UsageCoverageNotice no longer shows a 'totals are partial' message; it only renders after all environments reach a terminal state, covering failures, stale environments, and duplicate sources.
• The UsageSkeleton uses fixed bar heights (SKELETON_BAR_HEIGHTS) to avoid layout shifts and is replaced atomically when data is ready.

📊 Macroscope summarized e7aa04d. 1 file reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted

🗂️ Filtered Issues

No issues evaluated.

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c596fa78-a3ed-45f3-94bf-28f402e443b6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

🟡 Mediumusage/UsagePage.tsx:43

The settling gate only treats an environment as still reporting when summary is null, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, useUsage retains each environment's previous successful summary while marking its query as waiting, so every refreshed environment still counts as answered (isPending and isPartial both stay false) and settling is never set. The gate should also consider each EnvironmentUsageStatus's in-flight/waiting state — not just absence of a cached summary — so the skeleton shows until every refresh round-trip completes.

🚀 Reply "fix it for me" or copy this AI Prompt for your agent:

In file @apps/web/src/components/usage/UsagePage.tsx around line 43:
The `settling` gate only treats an environment as still reporting when `summary` is `null`, so clicking Refresh leaves the loaded totals visible and lets each environment's refreshed data pop in individually as responses arrive. During a refresh, `useUsage` retains each environment's previous successful `summary` while marking its query as waiting, so every refreshed environment still counts as answered (`isPending` and `isPartial` both stay false) and `settling` is never set. The gate should also consider each `EnvironmentUsageStatus`'s in-flight/waiting state — not just absence of a cached `summary` — so the skeleton shows until every refresh round-trip completes.

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Refresh skips settling gate

Medium Severity

settling only follows isPending / isPartial, which treat an environment as done once it has a non-null summary. Usage queries use SWR, so refresh keeps prior summaries while waiting is true. Hitting refresh therefore never re-enters the skeleton gate, and merged totals jump again as each device’s refresh lands—the same issue this change aims to stop.

^{Reviewed by Cursor Bugbot for commit e7aa04d. Configure here.}

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	11.3 KiB	11.3 KiB	−1 B (−0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	5.5 KiB	5.5 KiB	+7 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−8 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	49.7 KiB	49.7 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	11.4 KiB	11.3 KiB	−8 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	5.5 KiB	5.5 KiB	−5 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	5.9 KiB	5.9 KiB	−3 B (−0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	50.6 KiB	50.6 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: 288d8e3 · PR result: e7aa04d · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 94.6 KiB
• Claude decoded thread snapshot: 95.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

Approvability

Verdict: Needs human review

1 blocking correctness issue found. Two unresolved findings identify that the settling gate doesn't handle the refresh case - when refreshing, cached summaries are retained so the skeleton never shows, allowing totals to jump again. The Macroscope finding at Medium severity meets the repository's blocking threshold.

^{You can customize Macroscope's approvability policy. Learn more.}