Agent pane: other agents' messages draw as cards from their sender, not your prompt by teamleaderleo · Pull Request #22013 · manaflow-ai/cmux

teamleaderleo · 2026-10-19T06:46:37Z

When another agent messages a cmux NEXT chat, the pane drew the message as if the person had typed it. That covers acpmux's own deliveries (send_to_session from #21858 and cmux agent message, which carry _meta.provenance.kind: "agent_message"), and an adopted Claude Code session (acpmux records its history with client: "adopt") showed its history's <cross-session-message from=… from-name=…>, <task-notification> and <system-reminder> prompts as raw XML.

Now:

A message from another agent draws as the existing message card (MessageCard's look, cv-message), turned around: sender → This agent, two lines with Show more. With acpmux provenance the sender is a chat here, so the card has Open , which opens it beside this chat (chat.offer.open, the same path a started task's row uses).
A Claude Code <cross-session-message> shows its from-name (else its address) and the body only. Claude Code's trailing note to the agent is dropped from the card. Opening the card shows the whole prompt as the agent got it.
A <task-notification> draws as one quiet tool-row line with its summary. A prompt that's only reminders, local-command caveats or prompt-submit hook text gets no row. A typed prompt behind them draws without them.
These text envelopes are read only from adopted history. A live prompt's text is always the person's, so a client can't pass its words off as another agent's. Only acpmux's socket-only provenance can do that (prompt_provenance.rs).
A row keeps the sender's words, or the cleaned prompt, as its text, so copy and the minimap don't show raw XML.
A message from another agent doesn't come back through ↑ prompt recall. Task results (offer_result) are excluded from it now too.

No new strings: the card reuses message.label, message.self, message.more/less and subagents.open. Nothing new is gated, since the session tools ship on by default.

Parsing is in conversation/incomingMessage.ts. direct.ts sets row.incoming on a user_message row and App.tsx draws it.

Verification

conversation/incomingMessage.test.ts (bun): red on 7ccd9e8 (the module didn't exist yet), green on the fix. It covers provenance with and without fromName, a person's prompt that quotes an envelope (stays the person's), an unclosed envelope, entity decoding, and the notification and hidden cases.
test/agent-pane-suggested-tasks.test.ts gains a "messages from other agents" block in the real pane against the stand-in daemon, with the suggested-tasks toggle off: the provenance card with Open, the cross-session card with no raw XML and the envelope reachable when opened, and the notification and reminder cases. It's a browser-lane test, so it runs in CI only. I haven't run it locally.
tsc --noEmit and the existing conversation/ bun tests pass. verify-local.py --affected fails only at localization, with 34k parity errors in catalogs this PR doesn't touch (same on base).
A correctness review subagent found four issues: a reminder before a typed prompt hid the prompt, bookkeeping-only rows left an empty focusable gap, live prompts could forge a card, and numeric entities didn't decode. All four are fixed in 3d13829, with unit and browser cases added.
Also: send_to_session calls from the cmux-sessions MCP server (acpmux: agents start, list, read and message other sessions (cmux-sessions MCP) #21858) draw as message cards, not raw MCP rows (agentMessages.test.ts, red on 7c2f0e9).
Not dogfooded in the app. A queued agent message still shows in the composer queue as an ordinary queued prompt.

Changelog

Changed: cmux NEXT draws another agent's message to a chat as a card from that agent, not as your own prompt, and no longer shows Claude Code's raw message envelopes.

🤖 Generated with Claude Code

…prompts Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

A prompt acpmux delivered from another agent (agent_message provenance: send_to_session, cmux agent message) drew as the person's own bubble, and an adopted Claude session's <cross-session-message>, <task-notification> and <system-reminder> prompts drew as raw XML. They now draw as a message card from the sender (Open beside for a chat here; the raw envelope once opened), a quiet task line, or nothing, and stay out of the prompt history. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

coderabbitai · 2026-10-19T06:46:45Z

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

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: manaflow-ai/cmux/.coderabbit.yaml
Review profile: ASSERTIVE
Plan: Advanced
Run ID: ade7cc05-c899-4db5-9787-a279c9a22e72

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

Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

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

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

teamleaderleo · 2026-10-19T06:49:28Z

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

The cmux-sessions MCP tool from #21858 drew as a raw MCP tool row; it now draws as the SendMessage card does, to the chat it names. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…w for bookkeeping Review fixes: a reminder before a typed prompt hid the prompt; it now draws without the reminder. Bookkeeping alone gets no row instead of an empty, focusable 34px one. Envelopes are read only from adopted Claude history (client: adopt), so a live prompt can't pass itself off as another agent's. Rows keep the sender's words as their text, so copy and the minimap don't show raw XML. Numeric entities decode. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

cursor · 2026-10-19T06:55:08Z

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.

- vp fmt webviews/src/agent-session/acpmux/replyHref.ts - vp fmt webviews/src/pages/markdown/README.md - vp fmt webviews/src/pages/markdown/links.ts - vp fmt webviews/test/markdown-cmux-app-links.test.ts Autofix-Run: https://github.com/manaflow-ai/cmux/actions/runs/39212708861

github-actions · 2026-10-19T07:04:15Z

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

Review 8d37e25: 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=pr22013 -f ref=next-xsession-msgs -f platform=macos -f backend_mode=local

Unclear: web / react-apps-check (3) failed: no known signature; failed step: Test webviews. This PR merged before its CI finished: fix forward on main.

CI failed on 8d37e2558d (run 39508831122 attempt 1): 3 code, 2 unknown.

Job	Verdict	Why
web / react-apps-check (2)	code	a web test failed
web / react-apps-check (4)	code	a web test failed
web / react-apps-check (3)	unknown	no known signature; failed step: `Test webviews`
web / react-apps-check (1)	code	a Python test failed
web / react-apps-check	unknown	no known signature; failed step: `Check webview test shards`

Failures:

seen on other PRs categories.test.ts every page row has exactly one home: failed (job)
seen on other PRs editors.test.tsx editors > every schema row renders the editor for its kind: failed (job)
seen on other PRs new-tab-composer-still.test.ts chromium: the default page's composer stays put in a 900px window: failed (job)
seen on other PRs new-tab-composer-still.test.ts webkit: the default page's composer stays put in a 900px window: failed (job)
seen on other PRs new-tab-composer-still.test.ts chromium: the default page's composer stays put in a 560px window: failed (job)
seen on other PRs new-tab-composer-still.test.ts webkit: the default page's composer stays put in a 560px window: failed (job)
seen on other PRs agent-pane-no-capsules.test.ts agent pane page > rounds its controls as small-radius rects, not capsules: failed (job)
seen on other PRs agent-pane-question-records.test.ts answered live, the card goes and the record takes its place: failed (job)
seen on other PRs chat-find.test.ts Find in Chat > finds command output and edit diffs inside a closed Worked for fold: failed (job)
seen on other PRs new-tab-warm-first-frame.test.ts chromium: Cmd-T paints its final chats and project on its first frame: failed (job)
... 7 more in the job logs

Matched log lines

web / react-apps-check (2): (fail) every page row has exactly one home [1.10ms]
web / react-apps-check (4): (fail) agent pane page > rounds its controls as small-radius rects, not capsules [15.92ms]
web / react-apps-check (1): FAIL: test_coverage_base_gap_warns_on_a_pr_changing_the_gallery_format (__main__.ScopedChecksTests.test_coverage_base_gap_warns_on_a_pr_changing_the_gallery_format)

Not re-run automatically: web / react-apps-check (2), web / react-apps-check (4), web / react-apps-check (3), web / react-apps-check (1), web / react-apps-check are not machine failures.

Written by scripts/ci/classify_failures.py (ci-failure-attribution.yml); signatures are its SIGNATURES table. A machine verdict is the runner's fault, not this PR's; yours means the failing file is one this PR changes, also red on main that main's latest full suite fails the same way, seen on other PRs that it failed on another pull request's run lately.

red: cmux-next checks (god files, concurrency, crash safety, l10n); no finished base run on 2cd6ae1 to compare; do not merge until green; see lane-rules
red: web / Web status; NEW: green on the base 2cd6ae1; do not merge until green; see lane-rules
red: web / react-apps-check; no finished base run on 2cd6ae1 to compare; do not merge until green; see lane-rules
red: web / react-apps-check (1); no finished base run on 2cd6ae1 to compare; do not merge until green; see lane-rules
red: web / react-apps-check (2); no finished base run on 2cd6ae1 to compare; do not merge until green; see lane-rules
red: web / react-apps-check (3); no finished base run on 2cd6ae1 to compare; do not merge until green; see lane-rules
red: web / react-apps-check (4); no finished base run on 2cd6ae1 to compare; do not merge until green; see lane-rules

CI fast guards failed on 2c3c47fb7e (https://github.com/manaflow-ai/cmux/actions/runs/39457270393). It does not block the merge; a red guard merged into main breaks it for every open PR.
The log named no failed step; see the run.

Agents: python3 scripts/ci/guard_attribution.py fix applies the mechanical fixes locally.

…on-msgs # Conflicts: # webviews/src/agent-session/acpmux/replyHref.ts # webviews/src/pages/markdown/links.ts

…on-msgs # Conflicts: # webviews/src/agent-session/acpmux/App.tsx

greptile-apps · 2026-10-19T19:32:29Z

Confidence Score: 4/5

[Medium impact] The changes affect how messages and prompt history appear in the cmux NEXT agent pane.

Fix the sender’s Open action before merging.

Findings

Sender chat will not open ▶

Summary

This PR draws incoming agent messages as sender cards instead of user prompts.

Messages from other agents appear as cards from their sender.
Prompt recall skips messages and results the person did not type.
Messages sent with cmux-sessions tools appear as agent message cards.

_{Reviews (7) · Last reviewed commit: "Merge feat-cmux-next into next-xsession-..." · Reviewed by Greptile}

teamleaderleo · 2026-10-19T19:37:12Z

The react-apps-check red ("Check diff viewer eager JS budget") is a base red. On feat-cmux-next db370d0 the markdown page measures 1,407,890 of its 1,400,000-byte budget, the same as this branch. This PR changes no markdown-page code. The other web check failures on the earlier head (settings/icons.tsx formatting, CookieImportConsentPanel.swift focus ring, syncMotion.gallery.css raw font sizes) are in base files too.

…on-msgs # Conflicts: # webviews/src/agent-session/acpmux/model.ts

cursor · 2026-10-19T19:44:56Z

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.

teamleaderleo · 2026-10-19T20:18:03Z

Webview tests at head 4b38b4a, run on big-red. CI skips these because the base's markdown-page eager JS budget step fails before them.

CMUX_BROWSER_TESTS=1 bun test test/agent-pane-suggested-tasks.test.ts src/agent-session/acpmux/conversation/: 46 pass, 0 fail. This covers the three "messages from other agents" browser cases: the provenance card with the Open button, the adopted cross-session card with its raw envelope folded, and notifications and reminders with no raw XML.
bun run typecheck: no type errors in 1603 files.

Observed / Not run

Observed: the runs above. Not run: a native app dogfood. The change is webview-only.

…on-msgs

…ges and steer for incoming messages A Claude Code prompt that is only bookkeeping keeps an empty row that draws nothing, so the answer it brought stays its own turn instead of joining the previous prompt's. Retry on another agent's message sends the prompt as the agent got it, sender and envelope included. A task's result never comes back with the up arrow, whether suggested tasks show or not. Another agent's message card shows the images it sent and a steer's read state. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…r prompt Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

blacksmith-sh · 2026-10-20T02:39:39Z

Found 35 test failures on Blacksmith runners:

Failures

Test	View Logs
`/tmp/tmp9y_8nlzu/summary-missing.json (live):/summary-missing.json (live):`	View Logs
`/tmp/tmpvdfe45tm/summary-fail.json (live):/summary-fail.json (live):`	View Logs
`bundle-map.json does not require CEF's LICENSE.txt/` `bundle-map.json does not require CEF's LICENSE.txt`	View Logs
`bundle-map.json does not require CEF's LICENSE.txt/` `bundle-map.json does not require CEF's LICENSE.txt`	View Logs
`convention violations found (ERROR lines above)./` `convention violations found (ERROR lines above).`	View Logs
`src/pages/settings/categories.test.ts/categories.test`	View Logs
`src/pages/settings/categories.test.ts/categories.test`	View Logs
`src/pages/settings/editors.test.tsx/editors.test.tsx`	View Logs
`src/pages/settings/editors.test.tsx/editors.test.tsx`	View Logs
`test/agent-pane-no-capsules.test.ts/agent-pane-no-capsules.test`	View Logs

...and 25 more test failures. View all on Blacksmith: run 39457270393, run 39457270230

^{Need help on this PR? Tag @codesmith-bot with what you need.}

# Conflicts: # webviews/src/agent-session/acpmux/direct.ts

…ices Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

teamleaderleo · 2026-10-20T13:52:49Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Browser chrome icons is red on 11 of 11 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T13:53:06Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Capability texts never say update the app is red on 11 of 11 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

cursor · 2026-10-20T14:07:36Z

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.

teamleaderleo · 2026-10-20T14:12:47Z

CI / web / react-apps-check (3) is red on 9 of 15 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T14:22:34Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Checks job structure is red on 10 of 11 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T14:23:03Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Main-thread concurrency is red on 10 of 11 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T14:33:02Z

CI / web / react-apps-check (2) is red on 8 of 12 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T14:33:25Z

CI / web / react-apps-check (4) is red on 8 of 12 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T14:42:54Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › No process scanning (holders register themselves, plans/cmux-next/pty-holder.md 4) is red on 9 of 11 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T15:03:39Z

CI / web / react-apps-check (1) is red on 9 of 14 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T15:13:30Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Package conventions lint is red on 11 of 11 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

teamleaderleo · 2026-10-20T15:23:07Z

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Phone lane field allowlist is red on 10 of 10 fresh feat-cmux-next PR heads in the last 2:00:00, so it is most likely a base red, not this PR's. The base-reds lane has it; a base merge picks up the fix once it lands.

turns.ts keeps the hidden-bookkeeping skip and passes the base's sessionStatus; direct.ts imports both incomingPrompt and startedCardsShown; model.ts keeps the incoming/resend/untyped fields and the base's started field. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue

…reply Hunk by hunk: incomingPrompt keeps #22013's acp: sender id and rendered body and adds this branch's msgId/threadId; the card takes both the sent images and Reply; App passes images, steer and onReply; direct.ts takes #22013's row (hidden bookkeeping kept, untyped); conversation.css keeps both the Reply and steer rules; the gallery entry is #22013's plus a reply variant; Localizable.xcstrings is #22013's file plus the four message.reply* keys. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue

…ext-incoming-reply direct.ts keeps the reply's IncomingMessage import and prompt_accepted handler next to the base's startedCardsShown and harnesses_changed start mode refresh; Localizable.xcstrings is #22013's file plus the four message.reply* keys. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue

The base's incoming message files are #22013's squash unchanged, so each conflict keeps this branch's Reply edit of those lines. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue

…card fields #22013 opens and answers only a chat's cmux address (acp:<id>) and the card now carries the message's msgId and threadId, so the fixtures send from acp:child-3 and the incomingPrompt expectations include both ids. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue

github-actions · 2026-10-21T00:53:01Z

Shipped in:

cmux NEXT nightly 1.0.0-nightly.3951982493001 (df694bddc1)
cmux NEXT nightly 1.0.0-nightly.3952240756001 (1950034e60)

* Agent pane: test that other agents' messages draw as theirs, not raw prompts Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: other agents' messages draw as cards from their sender A prompt acpmux delivered from another agent (agent_message provenance: send_to_session, cmux agent message) drew as the person's own bubble, and an adopted Claude session's <cross-session-message>, <task-notification> and <system-reminder> prompts drew as raw XML. They now draw as a message card from the sender (Open beside for a chat here; the raw envelope once opened), a quiet task line, or nothing, and stay out of the prompt history. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: test that send_to_session calls draw as message cards Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: send_to_session calls draw as message cards The cmux-sessions MCP tool from #21858 drew as a raw MCP tool row; it now draws as the SendMessage card does, to the chat it names. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: read Claude Code envelopes only in adopted history; no row for bookkeeping Review fixes: a reminder before a typed prompt hid the prompt; it now draws without the reminder. Bookkeeping alone gets no row instead of an empty, focusable 34px one. Envelopes are read only from adopted Claude history (client: adopt), so a live prompt can't pass itself off as another agent's. Rows keep the sender's words as their text, so copy and the minimap don't show raw XML. Numeric entities decode. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * chore(next): PR autofix at 3d13829 - vp fmt webviews/src/agent-session/acpmux/replyHref.ts - vp fmt webviews/src/pages/markdown/README.md - vp fmt webviews/src/pages/markdown/links.ts - vp fmt webviews/test/markdown-cmux-app-links.test.ts Autofix-Run: https://github.com/manaflow-ai/cmux/actions/runs/39212708861 * Agent pane: Reply in place on another agent's message card An incoming agent message card (IncomingMessageCard, cx-eibkk) gets a Reply action (button, or R with the card focused). The answer goes to the sending chat as send_to_session sends a message: a session/prompt to that chat, attached hidden, with an agent_message provenance naming this chat, the message's threadId and replyTo the message, so it lands in the sender's chat as a threaded reply. acpmux took provenance only over the unix socket. The app pane (LocalApp) may now send one only as a reply: replyTo set, from another chat of this daemon, to the prompt's chat. Web and peers are still refused. The pane relay allows `provenance` in session/prompt's _meta.acpmux. Follow-up to cx-eibkk. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * acpmux: let the app pane's reply past the origin guard, only for a received message The remote guard refused any provenance from LocalApp before the reply rule ran. It now leaves LocalApp to prompt_provenance, and Hub::app_reply_fits requires the replying chat's log to hold the agent message (`replyTo`) the target chat sent it. The shared pane policy lists `provenance` as the Swift table does. A `cmux agent message` sender (`acp:<id>`) is answered by its id. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * acpmux: move the app reply check out of requests.rs (god-file budget); rustfmt Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: a reminder-only prompt keeps its turn; Retry, recall, images and steer for incoming messages A Claude Code prompt that is only bookkeeping keeps an empty row that draws nothing, so the answer it brought stays its own turn instead of joining the previous prompt's. Retry on another agent's message sends the prompt as the agent got it, sender and envelope included. A task's result never comes back with the up arrow, whether suggested tasks show or not. Another agent's message card shows the images it sent and a steer's read state. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: test that a reminder's turn offers no Retry of the earlier prompt Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: gallery entry for the incoming message card and its Reply Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * chore(next): PR autofix at c426d09 - webviews/src/agent-session/acpmux/conversation/IncomingMessageCard.tsx Autofix-Run: https://github.com/manaflow-ai/cmux/actions/runs/39359952105 * Agent pane: gallery entry for messages from other agents and task notices Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Fix agent message cards: open by session id, show the sender words A message cmux delivers to an acpmux chat carries the sender address as provenance from: acp:<session id> for a chat, term_... for a terminal agent, cli for a person at a shell. The card used it verbatim as the chat to open, so Open asked the host for acp:<id>. Strip acp: for the session id, and give no Open for any other address. The prompt text is cmux rendered message (agent_message_prompt.rs render): a header, a delivery note, ---, the body, and a closing line naming the message id. The card now shows the body, read only when the whole render matches and the closing line names provenance msgId, and keeps the full render as raw. Any other text shows whole. A task notification turn no longer offers Retry, which resent the notification XML as the person prompt. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(agent-pane): the reply tests follow #22013's sender address and card fields #22013 opens and answers only a chat's cmux address (acp:<id>) and the card now carries the message's msgId and threadId, so the fixtures send from acp:child-3 and the incomingPrompt expectations include both ids. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue * test(acpmux): an app pane reply signs only as a chat the pane shows, in its thread, under its name (red) The app pane's reply in place (Origin::LocalApp) is refused unless its from chat is one this connection shows (attached and not held unseen), its threadId is the answered message's thread, and its fromName is that chat's title, name or id. Today app_reply_fits checks only that from got the message, so the pane could sign as any chat that did. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue * fix(acpmux): an app pane reply signs only as a chat the pane shows, in its thread, under its name Hub::app_reply_fits is default deny for what the app pane signs: from must be a chat this connection shows (attached, not held unseen), the reply's threadId must be the answered message's thread (its threadId, else its msgId), and a fromName must be the chat's own title, name or id as the pane writes one. Before, any chat that got the message could be named as the sender under any name and thread. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue * style(acpmux): rustfmt the app reply check Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?harness=claude&session=c1a6da4a-9c34-4cea-864c-8b045b51ac46&host=huge-blue * Agent pane: Reply on the received-message line uses the type tokens; the catalog keeps base key order Follows the adoption review: - The Reply key hint and its sent/failed note use --text-caption and --text-detail instead of literal font sizes (check-type-scale). - The App.tsx props are one per line (oxfmt). - Localizable.xcstrings is base plus the four message.repl* keys, in base order, instead of a reordered rewrite. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Agent pane: Reply's test opens it with its button, not the removed R key; the sent note wraps Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * agent pane: keep the strings catalog in base order; add only the four Reply strings Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * agent pane: tool output under a message keeps base margins; the reply editor lines up with the body The base merge swapped the two margins. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * acpmux, agent pane: a reply in place arrives as a cmux agent message, linked back and signed by acpmux The pane sends from and to as acp:<id> addresses, as cmux agent message does, so the sender chat links the reply and can answer it; acpmux refuses a bare id. The text is wrapped as cmux renders a message (sender, id, the message it answers, the not-your- operator note). acpmux signs fromName with the replying chat own name instead of matching the pane one, so a stale title is never refused. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * agent pane: closing a message closes its reply and keeps the draft; reply state survives scrolling Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: cmux-autofix[bot] <cmux-autofix[bot]@users.noreply.github.com>

teamleaderleo and others added 2 commits October 18, 2026 23:37

Agent pane: test that other agents' messages draw as theirs, not raw …

7ccd9e8

…prompts Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Agent pane: test that send_to_session calls draw as message cards

7c2f0e9

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

teamleaderleo and others added 2 commits October 18, 2026 23:49

Agent pane: send_to_session calls draw as message cards

241711a

The cmux-sessions MCP tool from #21858 drew as a raw MCP tool row; it now draws as the SendMessage card does, to the chat it names. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

teamleaderleo added 2 commits October 19, 2026 12:25

Merge remote-tracking branch 'origin/feat-cmux-next' into next-xsessi…

b1f44dc

…on-msgs # Conflicts: # webviews/src/agent-session/acpmux/replyHref.ts # webviews/src/pages/markdown/links.ts

Merge remote-tracking branch 'origin/feat-cmux-next' into next-xsessi…

b949974

…on-msgs # Conflicts: # webviews/src/agent-session/acpmux/App.tsx

greptile-apps Bot reviewed Oct 19, 2026

View reviewed changes

Comment thread webviews/src/agent-session/acpmux/direct.ts Resolved

Comment thread webviews/src/agent-session/acpmux/composer/promptHistory.ts Resolved

Merge remote-tracking branch 'origin/feat-cmux-next' into next-xsessi…

4b38b4a

…on-msgs # Conflicts: # webviews/src/agent-session/acpmux/model.ts

This comment has been minimized.

Sign in to view

greptile-apps Bot reviewed Oct 19, 2026

View reviewed changes

Comment thread webviews/src/agent-session/acpmux/direct.ts Resolved

teamleaderleo mentioned this pull request Oct 19, 2026

gh-merge-green: excuse a red lint check whose every finding the base also prints #22128

Merged

Merge remote-tracking branch 'origin/feat-cmux-next' into next-xsessi…

6e023b2

…on-msgs

greptile-apps Bot reviewed Oct 19, 2026

View reviewed changes

Comment thread webviews/src/agent-session/acpmux/App.tsx Resolved

teamleaderleo and others added 3 commits October 19, 2026 19:15

Merge remote-tracking branch 'mf/feat-cmux-next' into pr-22013

79c2433

Agent pane: test that a reminder's turn offers no Retry of the earlie…

81fa415

…r prompt Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

teamleaderleo mentioned this pull request Oct 20, 2026

Agent pane: Reply in place on another agent's message card #22310

Merged

teamleaderleo and others added 2 commits October 19, 2026 19:53

Merge remote-tracking branch 'mf/feat-cmux-next' into pr-22013

0e9746d

# Conflicts: # webviews/src/agent-session/acpmux/direct.ts

Agent pane: gallery entry for messages from other agents and task not…

f14f221

…ices Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Merge branch 'feat-cmux-next' into next-xsession-msgs

5f48961

Merge branch 'feat-cmux-next' into next-xsession-msgs

e42e6bf

teamleaderleo added 2 commits October 20, 2026 07:44

Merge branch 'feat-cmux-next' into next-xsession-msgs

28b12e6

Merge branch 'feat-cmux-next' into next-xsession-msgs

5ab88c8

Merge branch 'feat-cmux-next' into next-xsession-msgs

6bd931c

Merge branch 'feat-cmux-next' into next-xsession-msgs

2c3c47f

teamleaderleo mentioned this pull request Oct 20, 2026

Agent pane: messages from other sessions draw as received lines #21747

Closed

3 of 4 tasks

teamleaderleo merged commit 793e1c0 into feat-cmux-next Oct 20, 2026

teamleaderleo deleted the next-xsession-msgs branch October 20, 2026 22:32

This was referenced Oct 21, 2026

cmux NEXT agent pane: a click on a sub-session opens its chat (chat.offer.open) #23159

Merged

cmux-next agent pane: Claude desktop polish gaps, ranked #21604

Open

Conversation

teamleaderleo commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Verification

Changelog

Uh oh!

coderabbitai Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review skipped

Uh oh!

teamleaderleo commented Oct 19, 2026

Uh oh!

cursor Bot commented Oct 19, 2026

Bugbot needs on-demand usage enabled

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

greptile-apps Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Confidence Score: 4/5

Findings

Summary

Uh oh!

Uh oh!

Uh oh!

teamleaderleo commented Oct 19, 2026

Uh oh!

cursor Bot commented Oct 19, 2026

Bugbot needs on-demand usage enabled

Uh oh!

This comment has been minimized.

Uh oh!

teamleaderleo commented Oct 19, 2026

Observed / Not run

Uh oh!

Uh oh!

blacksmith-sh Bot commented Oct 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Failures

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

cursor Bot commented Oct 20, 2026

Bugbot needs on-demand usage enabled

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

teamleaderleo commented Oct 20, 2026

Uh oh!

github-actions Bot commented Oct 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

teamleaderleo commented Oct 19, 2026 •

edited

Loading

coderabbitai Bot commented Oct 19, 2026 •

edited

Loading

github-actions Bot commented Oct 19, 2026 •

edited

Loading

greptile-apps Bot commented Oct 19, 2026 •

edited

Loading

blacksmith-sh Bot commented Oct 20, 2026 •

edited

Loading

github-actions Bot commented Oct 21, 2026 •

edited

Loading