fix(cmux-next): replay agent page events across bridge gaps by teamleaderleo · Pull Request #24617 · manaflow-ai/cmux

teamleaderleo · 2026-10-23T21:43:03Z

Summary

Agent-page host events are now buffered during a WebView bridge gap and replayed in order on reconnect. Replay-state events stay ahead of gap events, events arriving during subscription handoff cannot overtake the queue, and cancellation returns an undelivered suffix when no listener remains. The bounded queue applies backpressure instead of silently evicting events, and page-reset drops are requeued.

Leo said yes.

Performance evidence

The opt-in Swift harness uses 1,024 ordered events on the page-provider bridge and asserts all 1,024 are delivered in order. The same before/after benchmark was submitted at the final head, but the fleet is currently host-lock saturated:

baseline parent: earlier attempts 3dab501dc0cdad7327cc73d2 and e19... failed before timing because the benchmark source lacked its CmuxNextSettings import; that harness compile defect is fixed in this branch.
fixed head a2176e62e5fcf06d227de90cfefd933bbe990be4: final focused/benchmark submissions were cancelled after the branch changed; the fleet is currently host-lock saturated and produced no terminal benchmark receipt.
No publish_main_ms or replay_window_ms value is claimed until a terminal fleet receipt exists.

Validation

AgentPageProviderEventsTests covers replay ordering, fan-out, cancelled-replay recovery, reset-drop ordering, backpressure, reconnect clearing, and pending replay recovery.
Opt-in fixed 1,024-event harness: AgentPageProviderBridgeGapBench.
Independent Codex review of exact head a2176e62e5fcf06d227de90cfefd933bbe990be4: pending; prior exact-source review SHIP on the unchanged implementation before the base-only rebase.
Exact head: a2176e62e5fcf06d227de90cfefd933bbe990be4.
No new timers or observers were added.
No capture: bridge buffering/replay behavior only; no visual surface change.

Agent session: big-red

github-actions · 2026-10-23T21:43:16Z

cmux status _{(one comment, edited in place; each line is updated by the job that owns it)}

Review a2176e6: HQ feed card · gallery before/after runs · checks

App of this head (app.zip, kept 1 day)

gh workflow run reload-build.yml -R manaflow-ai/cmux --ref feat-cmux-next -f tag=pr24617 -f ref=perf-agent-chat-bridge-buffer -f platform=macos -f backend_mode=local

red: cmux-next checks (god files, concurrency, crash safety, l10n); no finished base run on 1e2573c to compare; do not merge until green; see lane-rules
red: cmux-next crash ratchet; NEW: green on the base 1e2573c; do not merge until green; see lane-rules

Tour of a2176e62e5 (run): no media. The tour job failed before the tour ran (build or setup); the run log has the output.

teamleaderleo · 2026-10-23T21:51:57Z

Bead cx-4b57c created for this PR by the cmux coordinator.

cursor · 2026-10-23T22:13:24Z

Bugbot needs on-demand usage enabled

Bugbot uses usage-based billing for this team and requires on-demand usage to be enabled.

A team admin can enable on-demand usage in the Cursor dashboard.

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a83-da67-7390-b026-b5d02463bfa0&host=big-red

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

github-actions · 2026-10-24T04:37:47Z

Shipped in:

cmux NEXT nightly 1.0.0-nightly.4005909948001 (66747a14cb)
cmux NEXT nightly 1.0.0-nightly.4005999011801 (09af538c30)

…hen one script Rebased from pl-scrolling d493af1 and 87895e9 onto the tip's router (the drop-and-replay sends of #24617), on the one send path and the visibility seam of the previous commits. PageRouter.deliver no longer sends an event to a page nobody can see (out of a window, hidden, parked, the pool's parked spare, or a covered or minimized window; automation launches keep covered windows live, as WindowDrawPolicy does for every surface). The events wait in order, with their drop actions, and go in one send when the page shows. Every event is kept (streams carry deltas). Anything else that reaches the page first releases what waits, so nothing reorders: a command and a popup event (the person's own action) go at once after it, the reply to a page message goes after it, a host call carries it ahead in the same send. Unsubscribe drops that subscription's waiting events; close drops them all and runs their drop actions after the script FIFO's, so a provider replays in send order. A busy hidden stream goes in batches of PageRouter.heldLimit (1024: a bound on memory and on scripts, not a measured number; such a batch is over the pipeline's inline budget, so it renders off the main actor). The agent pane opts out through its descriptor (PageDescriptor.holdsEventsWhileHidden false on .agent): its frame pacer already holds a hidden pane's frames and still sends the ones the page waits on (a reply, prompt_accepted, a turn's end). The flag is on the descriptor, so a pooled host rebound to another page takes that page's rule. debug.page state reports events_held and events_waiting. Guard: PageRouterHiddenPageTests (red in the previous commit: 100 sends while hidden), plus the batch limit and the descriptor opt-out cases. Not built or run here (code-only task): no compile-check, package test build or bench has run on this commit.

… the page after the reply; with none held the next event goes at once (red) #24617 (4abd404) inverted the early-events check in PageRouter.subscribe: when a provider sent its current state from inside `subscribe` (the icon picker's open session), the router released and dropped those events; when nothing was held, it queued the next events behind a main-actor task instead of sending them at once. Both tests fail on 71df513. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…the page after the reply; with none held the next event goes at once PageRouter.subscribe held a provider's events sent from inside `subscribe` and was to flush them after the reply. #24617 (4abd404) inverted the check when it added `eventsAfterSubscribeReply`: with events held it released and dropped them (a provider's current state never reached the page), and with none held it scheduled the flush task, so every event until that task ran waited behind it. The check is `early.isHolding` again; a provider whose events follow the reply never calls `hold`, so it needs no case of its own. Guard: PageRouterEarlyEventsTests (red at the previous commit). PageRouterHiddenPageTests, which emits right after subscribing, failed at its fix commit for the same cause. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…hen one script Rebased from pl-scrolling d493af1 and 87895e9 onto the tip's router (the drop-and-replay sends of #24617), on the one send path and the visibility seam of the previous commits. PageRouter.deliver no longer sends an event to a page nobody can see (out of a window, hidden, parked, the pool's parked spare, or a covered or minimized window; automation launches keep covered windows live, as WindowDrawPolicy does for every surface). The events wait in order, with their drop actions, and go in one send when the page shows. Every event is kept (streams carry deltas). Anything else that reaches the page first releases what waits, so nothing reorders: a command and a popup event (the person's own action) go at once after it, the reply to a page message goes after it, a host call carries it ahead in the same send. Unsubscribe drops that subscription's waiting events; close drops them all and runs their drop actions after the script FIFO's, so a provider replays in send order. A busy hidden stream goes in batches of PageRouter.heldLimit (1024: a bound on memory and on scripts, not a measured number; such a batch is over the pipeline's inline budget, so it renders off the main actor). The agent pane opts out through its descriptor (PageDescriptor.holdsEventsWhileHidden false on .agent): its frame pacer already holds a hidden pane's frames and still sends the ones the page waits on (a reply, prompt_accepted, a turn's end). The flag is on the descriptor, so a pooled host rebound to another page takes that page's rule. debug.page state reports events_held and events_waiting. Guard: PageRouterHiddenPageTests (red in the previous commit: 100 sends while hidden), plus the batch limit and the descriptor opt-out cases. Not built or run here (code-only task): no compile-check, package test build or bench has run on this commit.

… the page after the reply; with none held the next event goes at once (red) #24617 (4abd404) inverted the early-events check in PageRouter.subscribe: when a provider sent its current state from inside `subscribe` (the icon picker's open session), the router released and dropped those events; when nothing was held, it queued the next events behind a main-actor task instead of sending them at once. Both tests fail on 71df513. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…the page after the reply; with none held the next event goes at once PageRouter.subscribe held a provider's events sent from inside `subscribe` and was to flush them after the reply. #24617 (4abd404) inverted the check when it added `eventsAfterSubscribeReply`: with events held it released and dropped them (a provider's current state never reached the page), and with none held it scheduled the flush task, so every event until that task ran waited behind it. The check is `early.isHolding` again; a provider whose events follow the reply never calls `hold`, so it needs no case of its own. Guard: PageRouterEarlyEventsTests (red at the previous commit). PageRouterHiddenPageTests, which emits right after subscribing, failed at its fix commit for the same cause. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…hen one script Rebased from pl-scrolling d493af1 and 87895e9 onto the tip's router (the drop-and-replay sends of #24617), on the one send path and the visibility seam of the previous commits. PageRouter.deliver no longer sends an event to a page nobody can see (out of a window, hidden, parked, the pool's parked spare, or a covered or minimized window; automation launches keep covered windows live, as WindowDrawPolicy does for every surface). The events wait in order, with their drop actions, and go in one send when the page shows. Every event is kept (streams carry deltas). Anything else that reaches the page first releases what waits, so nothing reorders: a command and a popup event (the person's own action) go at once after it, the reply to a page message goes after it, a host call carries it ahead in the same send. Unsubscribe drops that subscription's waiting events; close drops them all and runs their drop actions after the script FIFO's, so a provider replays in send order. A busy hidden stream goes in batches of PageRouter.heldLimit (1024: a bound on memory and on scripts, not a measured number; such a batch is over the pipeline's inline budget, so it renders off the main actor). The agent pane opts out through its descriptor (PageDescriptor.holdsEventsWhileHidden false on .agent): its frame pacer already holds a hidden pane's frames and still sends the ones the page waits on (a reply, prompt_accepted, a turn's end). The flag is on the descriptor, so a pooled host rebound to another page takes that page's rule. debug.page state reports events_held and events_waiting. Guard: PageRouterHiddenPageTests (red in the previous commit: 100 sends while hidden), plus the batch limit and the descriptor opt-out cases. Not built or run here (code-only task): no compile-check, package test build or bench has run on this commit.

… the page after the reply; with none held the next event goes at once (red) #24617 (4abd404) inverted the early-events check in PageRouter.subscribe: when a provider sent its current state from inside `subscribe` (the icon picker's open session), the router released and dropped those events; when nothing was held, it queued the next events behind a main-actor task instead of sending them at once. Both tests fail on 71df513. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…the page after the reply; with none held the next event goes at once PageRouter.subscribe held a provider's events sent from inside `subscribe` and was to flush them after the reply. #24617 (4abd404) inverted the check when it added `eventsAfterSubscribeReply`: with events held it released and dropped them (a provider's current state never reached the page), and with none held it scheduled the flush task, so every event until that task ran waited behind it. The check is `early.isHolding` again; a provider whose events follow the reply never calls `hold`, so it needs no case of its own. Guard: PageRouterEarlyEventsTests (red at the previous commit). PageRouterHiddenPageTests, which emits right after subscribing, failed at its fix commit for the same cause. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…hen one script Rebased from pl-scrolling d493af1 and 87895e9 onto the tip's router (the drop-and-replay sends of #24617), on the one send path and the visibility seam of the previous commits. PageRouter.deliver no longer sends an event to a page nobody can see (out of a window, hidden, parked, the pool's parked spare, or a covered or minimized window; automation launches keep covered windows live, as WindowDrawPolicy does for every surface). The events wait in order, with their drop actions, and go in one send when the page shows. Every event is kept (streams carry deltas). Anything else that reaches the page first releases what waits, so nothing reorders: a command and a popup event (the person's own action) go at once after it, the reply to a page message goes after it, a host call carries it ahead in the same send. Unsubscribe drops that subscription's waiting events; close drops them all and runs their drop actions after the script FIFO's, so a provider replays in send order. A busy hidden stream goes in batches of PageRouter.heldLimit (1024: a bound on memory and on scripts, not a measured number; such a batch is over the pipeline's inline budget, so it renders off the main actor). The agent pane opts out through its descriptor (PageDescriptor.holdsEventsWhileHidden false on .agent): its frame pacer already holds a hidden pane's frames and still sends the ones the page waits on (a reply, prompt_accepted, a turn's end). The flag is on the descriptor, so a pooled host rebound to another page takes that page's rule. debug.page state reports events_held and events_waiting. Guard: PageRouterHiddenPageTests (red in the previous commit: 100 sends while hidden), plus the batch limit and the descriptor opt-out cases. Not built or run here (code-only task): no compile-check, package test build or bench has run on this commit.

…pt; later ones go at once #24617 swapped the two cases of the guard after provider.subscribe: when the provider sent its current state inside subscribe (the icon picker's open session), the held events were released and dropped; when it sent none, every later event waited in PageEarlyEvents for a flush task, a hop per subscription and out of step with the reply. The guard now flushes only when something is held and the provider does not send after the reply itself. Found by the PL-scrolling landing guard: PageRouterHiddenPageTests (every test that emits after the reply) went red on the rebased head. Tests: eventsSentDuringSubscribeFollowTheReplyAndLaterOnesGoAtOnce, eventsAfterAReplyWithNothingHeldGoAtOnce.

Each fails on the bare tip (full run at 2ee04b0; again alone at e869761, step 280e825d1b47e01c8df4b8ff). Stale premises (the test follows the change's intent): - IconFamilyAttributeTests: 5e7966c made folding finished thoughts the default; the test passes that default. - ContextMenuStandardTests: #24707 (0bf7675) put Inspect Element in the tab bar's + menu. - AcpmuxRefusedDaemonTests: #24795 (c7daff2) rewrote the refusal copy (no socket path; Quit that app; the holder-less refusal reconnects on its own) and left its own test on the old copy. - AgentPaneTranscriptStepTests, AgentPageProviderEventsTests: since #24617 (4abd404) a new subscription's events go out a main-actor turn later; the tests wait for them (bounded by the suite time limit). Real breaks (the code is fixed): - ActionSurfaceParityTests: #24707 made the sidebar section menu 13 top-level rows (the limit is 12); its Inspect Element goes in the Options folder. - TabSelectionActionsTests: Save and Close Tab (#23176) had a target reason for a menu row only, so the palette's run on a terminal tab was enabled and then refused; the reason now covers the untargeted run too. - SettingsSchemaExportTests: #24631 (78d7472) added four New Tab widget settings (clock, timer, warm greeting, quote) without strings or an agent policy: their 8 strings in all 21 locales, and the agent may set them. - NativeTooltipGuardTests: Mac Screen (ffc998b, 0372726) and No Color (c3bd2b2) drew raw tool tips: the No Color swatch uses themedToolTip, the toolbar's own helper is renamed setThemedHelp (it already set themedToolTip), the menu bar status item is marked native. - WorkspaceNameSourceTests: #24627 (f8c226e) counted any renamed '~' as unnamed (nameRevision > 0); only the automatic namer's '~' is. - AgentPaneMediaNoAutoActionTests: #24619 (5ff4b0c) resolved every image.load through the raster-only image roots, so a chat's own SVG or PDF was refused; non-raster files resolve in the chat's roots again. - BrowserProfileLinkMenuTests: e01bdd4 gave menu symbols their natural size; a profile's symbol is centered in the 16 pt square again, so names line up. - PhoneProcessRunner (passed alone, failed in the full run): a captured run ended before its stderr was read; it now waits for the end of stderr too. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…ge's push (#26778) Subscribing to the agent page host queues the replay and flushes it in a task (#24617), so a command published right after subscribe arrives on a later main-actor turn. eachStepSendsItsPageCommand and stopTurnSendsTheComposerStopCommand read the events synchronously and saw none (commands -> []). They now read the stream until the pushes arrive, as paletteSessionInfoReachesItsOwnPageCommand does since #25209. Agent-Session: https://cmux.com/deeplink/agent?id=aid-cc8bff1c3ddb4ec353c5 Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

#28492) (#28496) AgentPageProvider.subscribe sends a new subscription's events in a later main-actor turn, so they reach the page after the subscribe reply (#24617). routerResetDropUsesTheCapturedProviderAfterRouterCloses published right after router.handle(sub) and read the drop closure in the same turn, before the delivery. It has failed at AgentPageProviderEventsTests.swift:110 ('dropped' nil) in every cmux-next swift test run (jobs 124152289450, 124161816855, 124162143724, 124160068616) and at feat-cmux-next df9b52b. The test now waits for the router's delivery of that event (CountWaiter, as resetOfMultipleSubscribersRequeuesOnePublication does) before it checks that the drop requeues the event on the captured provider. No timing: it waits on the delivery itself.

…s the page (#27468) * test(cmux-next pages): a provider's state sent inside subscribe reaches the page (red) PageRouterEarlyEventsTests: a provider's state sent from inside subscribe reaches the page once, after the reply that names its subscription; an event after the reply goes to the page in its own turn. Both fail on feat-cmux-next 356f13f: #24617 inverted the guard in PageRouter.subscribe (fix in the next commit). Found on perf/chrome-parity, which has the same code (cp/fix-tip-page-script-batch-crash). * fix(cmux-next pages): a provider's state sent inside subscribe reaches the page PageRouter.subscribe keeps the events a provider sends while subscribe runs (its current state) and flushes them once after the reply (20066e1: `guard early.isHolding else { release }`). #24617 (4abd404) rewrote that guard for providers that send only after the reply and inverted it: - events held during subscribe were released and thrown away: the first state of the agent activity, icon picker and update surface pages (AgentActivityPageProvider, IconPickerProvider, UpdateSurfaceIsland call onEvent inside subscribe) never reached the page, which showed nothing until its next change; - with nothing held, events after the reply were caught and sent a main turn late. The guard is now `guard early.isHolding, !provider.eventsAfterSubscribeReply`. Regression test: PageRouterEarlyEventsTests (red in the previous commit). --------- Co-authored-by: lawrencecchen <>

teamleaderleo temporarily deployed to ci-routing-agent October 23, 2026 21:43 — with GitHub Actions Inactive

github-actions Bot added the risk:high Risk tier (docs/ci/risk-tiers.md) label Oct 23, 2026

This comment has been minimized.

Sign in to view

teamleaderleo temporarily deployed to ci-routing-agent October 23, 2026 21:52 — with GitHub Actions Inactive

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 21:57 — with GitHub Actions Error

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:00 — with GitHub Actions Error

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:03 — with GitHub Actions Error

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:04 — with GitHub Actions Error

teamleaderleo temporarily deployed to ci-routing-agent October 23, 2026 22:07 — with GitHub Actions Inactive

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:10 — with GitHub Actions Error

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:13 — with GitHub Actions Error

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:15 — with GitHub Actions Error

teamleaderleo temporarily deployed to ci-routing-agent October 23, 2026 22:16 — with GitHub Actions Inactive

teamleaderleo had a problem deploying to ci-routing-agent October 23, 2026 22:23 — with GitHub Actions Error

teamleaderleo temporarily deployed to ci-routing-agent October 23, 2026 22:24 — with GitHub Actions Inactive

teamleaderleo temporarily deployed to ci-routing-agent October 23, 2026 22:34 — with GitHub Actions Inactive

teamleaderleo added 11 commits October 23, 2026 16:26

fix(cmux-next): replay agent page events across bridge gaps

74a1158

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

test(cmux-next): benchmark agent page bridge-gap replay

e5f0e9f

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): bound agent page bridge replay

3f95bd0

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): replay dropped page bridge envelopes

2efdb60

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): retain provider for dropped page events

0c72504

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): signal bridge replay overflow

36cd016

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): close transport on page event overflow

0edf9bf

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): preserve reset event order

34eab12

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

fix(cmux-next): reset bridge overflow on transport reconnect

6b9b6fd

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a83-da67-7390-b026-b5d02463bfa0&host=big-red

fix(cmux-next): preserve partial bridge gap on reconnect

90def5c

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a83-da67-7390-b026-b5d02463bfa0&host=big-red

fix(cmux-next): reset stalled replay on reconnect

fc06345

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a83-da67-7390-b026-b5d02463bfa0&host=big-red

test(cmux-next): keep pending fanout regression bounded

625543e

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

teamleaderleo had a problem deploying to ci-routing-agent October 24, 2026 00:08 — with GitHub Actions Error

fix(cmux-next): order late bridge-gap recovery

a2176e6

Agent-Session: https://cmux.com/deeplink/agent?harness=codex&session=01a16a87-cdd1-7a41-ab1c-f59d9bf68cc8&host=big-red

teamleaderleo temporarily deployed to ci-routing-agent October 24, 2026 00:11 — with GitHub Actions Inactive

teamleaderleo merged commit 4abd404 into feat-cmux-next Oct 24, 2026
95 of 102 checks passed

teamleaderleo deleted the perf-agent-chat-bridge-buffer branch October 24, 2026 00:12

This was referenced Oct 24, 2026

fix(cmux-next): PageScriptPipeline compiles again (no covariant Self in a default argument) #24687

Closed

Mac Screen: stream at up to 120 fps when both displays can #24598

Merged

teamleaderleo mentioned this pull request Oct 28, 2026

test(next agent pane): transcript step and stop tests wait for the page's push #26778

Merged

Navigation Menu

Repository navigation

fix(cmux-next): replay agent page events across bridge gaps - #24617

fix(cmux-next): replay agent page events across bridge gaps#24617
teamleaderleo merged 34 commits into
feat-cmux-nextfrom
perf-agent-chat-bridge-buffer

teamleaderleo commented Oct 23, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Oct 23, 2026 •

edited

Loading

Uh oh!

This comment has been minimized.

teamleaderleo commented Oct 23, 2026

Uh oh!

cursor Bot commented Oct 23, 2026

Uh oh!

Uh oh!

github-actions Bot commented Oct 24, 2026 •

edited

Loading

Uh oh!

This branch was previously deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

teamleaderleo commented Oct 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Performance evidence

Validation

Uh oh!

github-actions Bot commented Oct 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

teamleaderleo commented Oct 23, 2026

Uh oh!

cursor Bot commented Oct 23, 2026

Bugbot needs on-demand usage enabled

Uh oh!

Uh oh!

github-actions Bot commented Oct 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This branch was previously deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

teamleaderleo commented Oct 23, 2026 •

edited

Loading

github-actions Bot commented Oct 23, 2026 •

edited

Loading

github-actions Bot commented Oct 24, 2026 •

edited

Loading