A gemini session stayed "working" for good after its first turn. Its braille
spinner trips the generic SPINNER_PATTERN and marks the pane working, but only
a composer glyph arms the idle confirmation and gemini declared none, so it
fell back to Claude's `❯`, which gemini never draws.
Measured on live Gemini CLI 0.63.0 panes (capture-pane every 300 ms through
real turns with a shell call at 40, 120 and 200 columns, YOLO and default
approval mode, plus the raw PTY stream). The turns ran against a local
stand-in for the Gemini API (GOOGLE_GEMINI_BASE_URL, which Codeman's custom
endpoint support already sets), since the CLI's TUI does not depend on the
backend and no account is needed for it:
- The TUI repaints its whole bottom region every frame, composer included,
and the composer sits between a `▄` bar and a `▀` bar; the submitted prompt
is echoed between the same bars. The `▀` bar arms the idle check: every
repaint carries it, tmux's reattach repaint too. The composer's prompt
character is no good: it follows the approval mode (`*` in YOLO), and its
`>` also starts the echoed prompt, which would make the submit verifier
read a submitted prompt as stranded and press Enter again.
- While a turn runs a line `⠦ Thinking... (esc to cancel, 6s)` animates about
every 80 ms (largest gap mid-turn: 214 ms). The label can be any loading
phrase, so the working line is the `(esc to cancel, <n>` suffix, or a
spinner frame opening a line for when a long phrase wraps that suffix.
Nothing at rest matches either.
- A tool confirmation (default mode) replaces the composer, stops the
spinner and the pane goes silent (3.9 s gap), so it reads as idle.
Verified on an isolated instance from this branch: a YOLO turn emitted one
session:working (+170 ms) and one session:idle (2.5 s after the last output);
a default-mode turn went working -> idle while the confirmation waited ->
working once allowed -> idle at the end; after a server restart four restored
gemini panes went busy -> idle in about 4 s; a fresh launch settled in 3 s.
The launch-settle and uncharacterised-CLI tests that used gemini as their
example of a CLI without work detection now use grok and deepseek.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A codex session waiting on a background terminal read as plainly idle,
with no "1 background terminal" badge. Codex pins that row above its
composer, and the registry looked for it in the last three non-blank
rows. Codex 0.162.0 added a hint row under the status line at rest
(` ← for agents · ? for shortcuts`), which pushes the chip to FOURTH from
the bottom exactly when the idle probe reads it. Measured live on the
1.36.0 beta with `sleep 600` started as a background terminal: chip,
composer, status line, hint row; the server reported `watching: null`.
While a prompt is typed the hint goes away and the chip is third again.
The codex entry now declares `watchingLines: 4`. The trade is stated in
the entry: with no terminal running, the fourth row from the bottom is
the last transcript row (the last two while typing), which the agent
writes. As before, this is contained by codex having no hooks: a forged
row costs a wrong badge, never a silenced alert.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- claude watchingLine: the lookahead keys on "Artifact" alone, so a
footer truncated mid-chip ("1 Artifact…", "1 Artifact comm…") is still
refused instead of reporting the shell beside it; comment follows
- test: both truncations return no watching label
- invariants: a chip that waits on a human never counts as watching, and
the ^ anchor is what stops the retry past the chip
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
An agent that publishes an artifact arms a monitor for its comments and
ends its turn. Claude Code shows that on the footer as `1 Artifact
comment monitor`, and #473 put that chip on the list of background work,
so the session counted as watching and its idle prompt opened already
acknowledged. Unlike every other chip on the list, that monitor waits
on the user: the agent hears nothing until somebody comments.
Claude's `watchingLine` now refuses any footer that carries the chip,
through a lookahead over the whole row, so a shell running beside the
monitor cannot report the session as watching either.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- session.ts: a pane capture that fails now CLEARS the watching label
(and emits watchingChanged so pages drop the badge) instead of keeping
the last one, so a failed capture degrades toward an alert rather than
pre-acknowledging the next real idle prompt. Test updated; invariant
noted in architecture-invariants.
- approvals-ui.js: the header bell counts only unacknowledged items
(pendingApprovalsCount), matching codeman tui's pendingApprovalCount();
pinned in watching-no-alert.test.ts.
- mobile-overview.js: move the orphaned "Pill copy per state" JSDoc back
onto MOBILE_OVERVIEW_PILL_LABEL.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Reported from a manual test: a Codex session went on showing the watching badge
after its background terminal had finished. The server was right and the page
was stale — `Session.watching` changes while the session's status does not, and
nothing broadcast it.
The label is usually SET on the idle transition, which broadcasts anyway, so the
badge always appeared correctly. It CLEARS when the work ends, and a CLI can end
background work without taking a turn: codex repaints its background-terminal
row away and stays idle, so `_confirmIdle()` concludes without emitting `idle`
(that emit is guarded by `wasWorking || isInitialReady`) and no other event
fires. Every open page kept drawing a badge the server had already dropped.
`_readWatching()` now emits `watchingChanged` when, and only when, the label
really changes, and the wiring pushes the session state on it. No new SSE event:
the badge reads off the session payload every surface already has.
A/B measured on an isolated beta with the page loaded and then left untouched.
Without this commit the server dropped the label at t+50s and the page still
showed the badge at t+100s; with it, page and server cleared in the same
ten-second window.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A dual review (Codex CLI and Claude's code-reviewer, same diff, same brief)
found the trust boundary weaker than the comments around it claimed. Eleven
findings, all applied.
The two blockers were both about who can write the row the label is read from.
Claude's window covered two rows, and the second one is the status line, whose
command a session running with permissions bypassed can write into its own
`.claude/settings.json` — so an agent could print `· 1 monitor ·` onto a row of
its own and silence its own idle alert. The default window is one row now, which
is the footer and nothing else, and the constant says why. Separately, the label
reached `data-tab-meta-sig` unescaped while the row is installed with innerHTML,
which is an injection sink for any config-supplied pattern whose capture group is
permissive; it goes through escapeHtml() like every other untrusted string in
that file.
The Codex entry could not be fixed the same way, and now says so. Its row is
third from the bottom only while a terminal runs; with none running that slot
holds the last row of the transcript, so matching the complete row (with the
`/stop to close` tail, window narrowed to three) raises the bar without closing
it. What contains it is `hooks: 'none'`: no hook event from a codex session
reaches notePrompt(), so a forged label costs a wrong badge and cannot quiet an
alert. The registry comment, `docs/cli-registry.md` and the test all state that
rather than claiming a guarantee the code does not have.
Also from the review: the TUI header badge no longer counts an acknowledged
item, which was the same gate the classifier fix already went through and was
wrong for human acknowledgement too; the TUI approval card reads the quiet
reason and drops to a new `info` tone instead of asking for a reply; the badge
carries an aria-label, because the phone it was built for has no hover target;
the schema refuses `watchingLines` without a `watchingLine`; and the pattern and
its window are resolved together rather than one memoized and one not.
Documentation moved with it. The mechanism now lives in
`docs/architecture-invariants.md` with CLAUDE.md keeping the rule and a pointer,
`docs/wiki/Notifications-And-Approvals.md` tells users why a session stopped
buzzing, and both that page and the changeset name the limitation neither did
before: a question asked in plain prose is not a dialog, so it is silenced along
with the false alarms while background work runs.
Verified live again after the narrowing, on an isolated beta: a Claude session
reported `1 monitor` and took its idle prompt acknowledged, and a Codex session
reported `1 background terminal` against the full-row anchor.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Codex states background work too, and it says so in a different place. Claude
writes `· 1 monitor ·` on the last row of the screen; Codex pins
`1 background terminal running · /ps to view · /stop to close` ABOVE its
composer, which puts that row third from the bottom once the status line and
the composer are counted.
So how far up the screen to look is now per-CLI data as well:
`capabilities.workDetect.watchingLines`, bounded to 1..8 by the schema, and
defaulting to Claude's two. That bound is the point. The window is half the
injection guard, since every row it adds is another row the agent itself may be
able to write, and the label is what silences an idle alert. The other half is
the anchor, and Codex's is ` · /ps to view`: chrome naming a slash command only
the CLI can offer, so a session that writes "I left 1 background terminal
running for you" into its own output matches nothing.
Measured against a live codex-cli 0.154.0 pane rather than read out of a
binary. The row appears when the terminal starts, follows the composer down as
the conversation grows, and is gone after `/stop`. Verified end to end on an
isolated beta: the session payload carried `watching: "1 background terminal"`
and the badge rendered with it, and both cleared when the terminal stopped. The
fixtures in the tests are that capture verbatim.
Codex has no hook signals, so no idle prompt and no false NEEDS YOU row: for a
Codex session this is the badge alone, which is the case the maintainer said a
registry field could cover and a hook never could. Cross-CLI tests pin that
neither pattern fires on the other's screen, and that a CLI declaring nothing
still reports nothing.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The badge alone left the row in NEEDS YOU, which is the thing the issue was
about. The fix is the alert that does not fire.
An idle prompt from a session that is watching its own background work now
opens ALREADY acknowledged. `hook-event-routes` passes `Session.watching` to
`notePrompt()`, which sets `acknowledgedAt` and records why in a new
`acknowledgedReason`. Nothing new suppresses anything: `acknowledge()` has
always meant "the alert this prompt armed is spent", and the prompt itself
stays pending, answerable and available as Read My Mind context. A wrong label
therefore costs a card that does not blink, never an alert that was never
created.
Every surface follows from that. The broadcast carries the reason, so a live
page declines to arm the tab alert and raises no desktop notification. The push
is skipped, since a false alarm is hardest to ignore on a phone. A reloading
page reads `acknowledgedAt` in `seedApprovals()`, which it already did. And
`classifySession()` now reads it too, which is a pre-existing bug fixed here:
acknowledging on one device cleared the alert everywhere except `codeman tui`.
It re-arms for free, because the next idle prompt supersedes the item and is
built fresh. Only `idle` is eligible, so a dialog that blocks the agent still
goes red whatever else it started.
The label is pane-derived and therefore prompt-injectable, so it is now read
from the last two rows of the screen only, with Claude's pattern anchored on
the `·` its footer joins items with, ANSI-stripped and length-capped at the
source. An agent that prints `· 1 monitor ·` into its own output finds no
match.
Verified on an isolated beta: a session that armed a monitor took its idle
prompt acknowledged with no alert on any surface, wore the badge, and showed
"quiet, watching 1 monitor" on its still-answerable card; the same session with
the monitor killed alerted normally on the next prompt. `test/watching-no-alert.test.ts`
pins both directions across all four surfaces.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
An agent that arms a monitor, backgrounds a shell or hands a task to a
cloud session is told to end its turn. The pane then falls quiet, Claude
Code's idle_prompt notification arrives a minute later, and every surface
files the session under NEEDS YOU with nothing for a human to answer.
Claude states what it is still running on the last row of its screen
(`⏵⏵ bypass permissions on · 1 monitor · ← for agents`). That row is now
`capabilities.workDetect.watchingLine` in the CLI registry, guarded by
compileVersionRegex() like every other config regex, and the idle probe
reads it off the capture it already takes: `watchingLabel()` in
session-activity.ts searches the last five lines only, so a session that
PRINTS "1 monitor" is not mistaken for one running it.
The label lands on Session.watching and rides toLightDetailedState() out
to every surface. The phone overview, the desktop home rail and the rich
sidebar rows wear it as a `watching` badge in the accent colour, beside
the state pill and never in place of it: an agent can arm a monitor and
ask a question in the same breath, and only the pill says which.
Verified end to end against a throwaway session on an isolated beta
instance: the payload carried `watching: "1 monitor"` once the turn
ended, the badge rendered next to a yellow `waiting` pill, and both
cleared when the monitor died.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Codeman maintainer