[teamleaderleo]

Leo's ask, his words: "ask user question tool and its various states need looking at allegedly there's some sort of vibe-coded thing that agents can do to have prettier ask question situations...? idk but we could look into that/research that then support it..."

Leo-ask: ask-question-states

What was wrong

Claude Code's AskUserQuestion lets an agent attach a preview (a mockup or code sample) to each option. The answer the host sends back may carry annotations (question text -> {preview?, notes?}), and the agent then reads "Q"="A" selected preview: <mockup>, so it knows which mockup the person picked. cmux sent only answers. The card never sent the chosen preview, the pane relay refused any annotations param, and acpmux would have dropped it anyway.

The fix

• Card (question/model.ts reply): for a Claude Code ask, each single-select item whose chosen option has a preview adds annotations[prompt] = {preview}. Other answers, multi-selects, the Chief and Codex send nothing new. App.tsx and direct.ts pass it through to _acpmux/permission_respond.
• Pane relay, in both Swift AcpmuxPaneMethods and cmux-agent-pane-policy, from the same policy.json: annotations is a known param. It goes only beside answers, to a pending question the daemon sent, keyed by that question's own items, and each value is {preview?, notes?} strings of at most 64 KiB.
• acpmux (questions::check_annotations, respond_permission_annotated):
  • It takes annotations only from the person's own connection, and only with answers. A phone's signed answer or a session's answer can't carry them.
  • For Claude Code's AskUserQuestion, a preview must match one of that question's option previews as normalized, and the agent gets its own original text back, since Claude Code refuses a preview it didn't write. Notes must be 1 to 8192 bytes. Any other field, or any other question, is refused.
  • For any other question, the annotations are dropped, so the answer still goes through.
• Protocol IR, generated TS types and acpmux-schema.json document the param.

Not in this PR: notes UI (the protocol carries notes now; the card doesn't offer a field yet). HTML previews, markdown descriptions, answered and skipped summaries, and MCP elicitation belong to the transcript-parity lane.

Relay analysis

This adds one param to an existing method of the agent pane relay (`_acpmux/permission_respond`). It doesn't add a method, and it changes nothing remote: `RemoteAcpmuxWire` still passes only the option.

• Command or content execution: none. The value is checked text that ends up in Claude Code's own tool input, and Claude Code validates it again (`admitCardAnswer`).
• Unowned objects: the relay accepts it only for a question the daemon sent this pane, keyed by that question's items. acpmux accepts it only from the person's connection, for the pending prompt, beside answers.
• Local-state exposure: none. Nothing is read back. A preview has to be text the agent itself wrote, and notes are capped at 8192 bytes.

Test

Red at the test-only commit 3fc192c, green at the fix:

• bun test src/agent-session/acpmux/question/model.test.ts src/agent-session/acpmux/question/record.test.ts (from webviews/, on cmux16): red 11 pass / 1 fail ("a Claude answer sends the chosen option's preview"), green 12 pass / 0 fail.
• acpmux questions_annotations* (tests/hub_integration/questions.rs: carried with the agent's own preview text; refused when the preview isn't the agent's, the key or field is unknown, notes are empty or too long, or answers are missing; dropped for a question that isn't Claude's) and the policy cases in tests/cases/check.json run on the fleet with cmux-tui-rust-check.sh test. Results are below.
• cmux-ci compile-check --app at the head: see below.

Observed / Not run

Observed: the webview tests above on cmux16. Also at the head on cmux16: question/, needsYou/ and test/question-card-submit.test.tsx all green, tsc --noEmit clean, and prepush-gate.py reports no new violations.

Not run: a live Claude Code run that reads the annotation back (the fake agent echoes the tool input instead), and the Swift relay unit tests (the shared policy cases cover the same rule in Rust; the Swift harness doesn't read questions state yet).

No capture: no visible UI change; the card looks the same and sends one more field.

Gallery: agent-pane.question-card

Changelog

Changed: When you answer an agent's question that has option previews, Claude Code now learns which preview you picked.

Adopted and rebased (2026-10-25)

The stale-pr-reconciler, word for word: "These PRs in your area have no live owner (Leo, 2026-10-25: "these should never have been this stale for this long"). Adopt each: pr-owner claim N (in that repo), rebase onto its base, an independent review at the new head, then land it".

• Squashed onto feat-cmux-next (it was 3,151 commits behind). Previews are now {text, format} objects on base; the card already reads preview.text, and check_annotations compares against the trimmed raw option text, so the agent still gets its own text back. The only conflict was where the new function sits in hub/questions.rs.
• Codex review (gpt-6-astra xhigh) round 1, BLOCK: 1. a selected preview over 65,536 UTF-8 bytes rode along as an annotation, and both pane relays then refused the whole answer. Fixed: the card leaves out a preview over the relay's bound (MAX_ANNOTATION_BYTES, the same as policy.json maximum_annotation_bytes), and the answer still goes through. Test first: a preview over the relay's 65,536-byte bound is left out, and the answer still goes (32,768 x "é" is exactly the bound; one byte more, and 65,537 ASCII bytes, are left out).
• Round 2: SHIP at f7c5b23. Round 3 (a vp fmt reformat of that test only): SHIP at 079ae3e.
• Rebased again onto 714af0c, past acpmux: chat multiplayer A: person author, provenance, answeredBy, queue.not_author #25268 (chat multiplayer: author, answeredBy). respond_permission_annotated now returns (pending, answered_by, annotations), and the request handler keeps base's answerer.author.
• Round 4 (fresh reviewer), BLOCK: 1. the annotation checks ran before the pending lookup, so a losing annotated answer from a non-person connection got person_required instead of -32002 with the first answer's answeredBy/outcome. Fixed by moving them to after the lookup. Test first: a_losing_answer_is_told_who_answered_first now also sends an annotated losing answer (red at 2eddd47, green at 2551dc9, hub_integration questions/permission/author 40 pass).
• Round 5: SHIP at 2551dc9. The Release compile (Xcode 26) passed at that head. The Swift test reds are in untouched areas (password manager route, icon family, sidebar rail), and the generated-files red is base's open-remote CLI name (cmux-next: one verb to open a persistent remote terminal (Linux VMs too): "open <host> in a bottom split over ssh" #25092).

Tests on the rebased branch (fast-checks on Blacksmith)

• Red at the test commit 2838a9b: questions_annotations_carry... and ..._refused_and_the_ask_stays fail; policy full_check_order fails on the new cases.
• Green at fb4cc7c: acpmux --test hub_integration questions 8 pass, --lib questions 5 pass, cmux-agent-pane-policy parity 16 pass; clippy -D warnings and fmt clean. cmux-agent-pane-policy --test typescript the_allowlist_is_what_the_page_sends fails the same way on base de67de9 (_acpmux/mcp/reconnect), so it isn't from this PR.
• Webviews: red at a898660 (the bound test fails), green at f7c5b23: question/, needsYou/, test/question-card-submit.test.tsx 46 pass, tsc --noEmit clean.
• Base reds this PR shows too: Script tests (release-compile-derived-data.test.sh "bench OTHER_SWIFT_FLAGS", fails on base 5e6cb13); gallery health agent-pane.composer#file-not-attached (broken on other PRs too).

No timers or observers added. No capture: the card looks the same and sends one more field.

Agent session: big-red
🤖 Generated with Claude Code

[github-actions]

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

Review 079ae3e: 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=pr24356 -f ref=aff-ask-annotations -f platform=macos -f backend_mode=local

Gallery before/after for 2551dc9: 16 states rendered, 0 images differ, 5 more touched states not rendered (picked by the Gallery: line in the PR body). Images are in the run's artifact, not on this public PR.

gh run download 40279999965 -R manaflow-ai/cmux -n gallery-advisory-24356-2551dc935a6a3db410425a083d3f7d0bf9eaa1d0 (then read gallery-review.json)

States

State	Theme	Status	Diff	Picked by
agent-pane.question-card/pending-single	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-single	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/folder-trust	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/folder-trust	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/folder-trust-picked	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/folder-trust-picked	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/pending-multi	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-multi	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/pending-other-typing	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-other-typing	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/pending-with-preview	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-with-preview	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/pending-with-preview-hover	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-with-preview-hover	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/pending-with-preview-fenced	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-with-preview-fenced	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/pending-4-questions	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/pending-4-questions	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/codex-user-input	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/codex-user-input	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/acp-interactive	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/acp-interactive	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/chief-asks	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/chief-asks	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/reply-refused	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/reply-refused	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/answered	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/answered	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/answered-remote	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/answered-remote	Apple System Colors	unchanged	0%	pr-body
agent-pane.question-card/answered-by-session	Apple System Colors Light	unchanged	0%	pr-body
agent-pane.question-card/answered-by-session	Apple System Colors	unchanged	0%	pr-body

_{Pick other states with a Gallery: entry.id, entry.id/variant line in the PR body.}

Probably yours: accent-glass-invariants.test.ts fails here but not on main's latest full suite. Seen on other PRs too (likely flaky): claimPaint.test.tsx, claimPaint.test.tsx, agent-pane-agents-pill.test.ts and 12 more.

CI failed on 079ae3ee5a (run 40263423131 attempt 1): 2 code, 1 unknown.

Job	Verdict	Why
web / react-apps-base-check (4)	code	a web test failed
web / react-apps-base-check (1)	code	a web test failed
web / react-apps-check	unknown	no known signature; failed step: Check webview test shards

Failures:

• seen on other PRs claimPaint.test.tsx a pooled Settings page claimed by the host reports its paint once its first screen is in: failed (job)
• seen on other PRs claimPaint.test.tsx a Settings page loaded fresh reports its first paint once its first screen is in: failed (job)
• new in this PR accent-glass-invariants.test.ts no orange or system accent and no opaque glass on 2140 gallery stages: failed (job)
• seen on other PRs agent-pane-agents-pill.test.ts by default the running agents sit in the tray above the composer, not in the transcript: failed (job)
• seen on other PRs agent-pane-step-rows.test.ts every step is one 26px row on one rhythm, across transcript rows: failed (job)
• seen on other PRs chat-environment-scroll-stability.test.tsx long-chat scroll leaves the Left chip mounted and stable, while own cwd changes update it: failed (job)
• seen on other PRs chat-environment-scroll-stability.test.tsx long-chat scroll leaves the Right chip mounted and stable, while own cwd changes update it: failed (job)
• seen on other PRs chat-inspector-overlay-toggle.test.ts chromium, header: the toggle is above the open overlay and a click on it closes the Inspector: failed (job)
• seen on other PRs agent-pane-background-tasks.test.ts the summary card lists the chat's background tasks: running ones first, the finished folded: failed (job)
• seen on other PRs agent-pane-hairline-focus.test.ts composer at 1x: focus is the resting hairline a step brighter, no ring: failed (job)
• ... 6 more in the job logs

Matched log lines

web / react-apps-base-check (4): (fail) a pooled Settings page claimed by the host reports its paint once its first screen is in [175.97ms]
web / react-apps-base-check (1): (fail) the summary card lists the chat's background tasks: running ones first, the finished folded [941.92ms]

Not re-run automatically: web / react-apps-base-check (4), web / react-apps-base-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 714af0c to compare; do not merge until green; see lane-rules

[blacksmith-sh]

Found 45 test failures on Blacksmith runners:

Failures

Test	View Logs
/tmp/tmpxhvgec95/summary-missing.json (live):/summary-missing.json (live):	View Logs
/tmp/tmpyxk26ofz/summary-fail.json (live):/summary-fail.json (live):	View Logs
[new_workspaces] debug.new_workspace timed 0 of 50 new workspace rows/ [new_workspaces] debug.new_workspace timed 0 of 50 new workspace rows	View Logs
[new_workspaces] the window draws no frames/[new_workspaces] the window draws no frames	View Logs
[tabs] debug.tab_switch timed 0 of 40 switches/ [tabs] debug.tab_switch timed 0 of 40 switches	View Logs
[tabs] the window draws no frames/[tabs] the window draws no frames	View Logs
bench OTHER_SWIFT_FLAGS is 'OTHER_SWIFT_FLAGS=$(inherited) -DCMUX_BENCH_HOOKS -DCMUX_BL OCKING_CHECKS'/ bench OTHER_SWIFT_FLAGS is 'OTHER_SWIFT_FLAGS=$(inherited) -DCMUX_BENCH_HOOKS -DCMUX_BL OCKING_CHECKS'	View Logs
bench OTHER_SWIFT_FLAGS is 'OTHER_SWIFT_FLAGS=$(inherited) -DCMUX_BENCH_HOOKS -DCMUX_BL OCKING_CHECKS'/ bench OTHER_SWIFT_FLAGS is 'OTHER_SWIFT_FLAGS=$(inherited) -DCMUX_BENCH_HOOKS -DCMUX_BL OCKING_CHECKS'	View Logs
convention violations found (ERROR lines above)./ convention violations found (ERROR lines above).	View Logs
test/agent-pane-background-tasks.test.ts/agent-pane-background-tasks.test	View Logs

...and 35 more test failures. View all on Blacksmith

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

[teamleaderleo]

Bead cx-ix34d created for this PR by the cmux coordinator.

[teamleaderleo]

Orphaned: no activity for 13 h, and no lane holds a lease on it. Adopted by the orphan sweep: merged feat-cmux-next in at 96079b6. A later pass queues it for the merge hub if it is reviewed, small and red only on base reds.

[teamleaderleo]

cmux-next crash-class rules / cmux-next crash-class rules is red on 13 of 13 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]

cmux-next guards / nightly-guards / workflow-guard-tests / release-notary is red on 17 of 17 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]

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Checks job structure is red on 17 of 17 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]

Owned by hb-acpmux on big-red · session

[teamleaderleo]

cmux-next fast gates / post cmux-next/fast-gates is red on 16 of 18 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]

high-signal review link / link is red on 14 of 18 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]

cmux-next / cmux-next checks (god files, concurrency, crash safety, l10n) › Script tests (pin-cmux-tui, embed-cef, ensure-cef store) is red on 7 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.

[github-actions]

Shipped in:

• cmux NEXT nightly 1.0.0-nightly.4030749448501 (b781e3b4a0)
• cmux NEXT nightly 1.0.0-nightly.4031371403702 (1dc93b2657)