Every run mode is now a `CliEntry` in `src/config/cli-registry/` — discovery
(search dirs, version + identity probes), the launch argv template, env
handling, the `capabilities` flags that replace per-CLI branching, and the
`overlays` that back the remote/docker pane commands. Code that used to ask
"which CLI is this?" reads the entry instead.
Behaviour is unchanged. `test/cli-registry-spawn-golden.test.ts` pins every
spawn command as a literal string, captured from the hand-written builders
before they were deleted, and `test/location-overlay-commands.test.ts` does the
same for all 20 remote and in-container pane commands.
Config can never contain shell text: an entry declares typed argv tokens,
literals are validated against a safe-word pattern at LOAD time (a bad literal
rejects the whole entry — a silently dropped `--no-approve` is not cosmetic),
and values resolve through patterns NAMED in code, so a user `clis.json` cannot
widen its own validation. `~/.codeman/clis.json` overrides any entry, read-only
in this release.
OMP is included as a registry entry rather than a tenth hand-written builder,
so `buildOmpCommand()`, the omp availability pre-flight, the omp arm of
`buildPathExport()` and the omp entries in the truecolor/NO_COLOR, alt-screen
and doctor ladders all drop out.
Guard rails:
- `test/cli-registry-no-id-branching.test.ts` fails the build if per-CLI-id
branching reappears outside `stock.ts`, in any of its four shapes (`===`,
`!==`, `switch`/`case`, `includes`) — an `===`-only version would miss the
negated forms, which is how 36 of them survived an earlier pass. Every
allowlisted branch carries its reason.
- `external`, `hooks` and `altScreen` stay three INDEPENDENT capabilities;
deriving one from another shipped the `until=stop`-hangs-on-shell bug.
- `param` is two namespaces. `launch.params` keys, `configSetenv.fromParam` and
`privilegedParams[].param` all name a LAUNCH param; the legacy `<Mode>Config`
wire field is separate, bridged only by `legacyConfigAliases`. Getting
`privilegedParams[].param` wrong is SILENT — it is the multi-user bypass
clamp's only handle on a CLI's privilege switch, and a wrong name clamps
nothing with no error and no failing test — so `schema.ts` rejects an entry
naming a param it never declared.
- Registry data resolves AT CALL TIME (`sessionModeSchema()`,
`allowedEnvPrefixes()`, `dependencyRegistry()`, the resolvers' `searchDirs`
thunks). A module-level const freezes at first import, so a CLI enabled while
the server ran moved the run menu but not that surface.
- Six fields are annotated DECLARED-FOR-LATER and read by nothing
(`shortBadge`, `accent`, `capabilities.echo`/`wheelForward`/
`keyboardAccessory`/`maxFrameBytes`): all frontend behaviour, transcribed
rather than measured. A test pins the list so it cannot quietly grow.
Three user-visible changes, all deliberate and named:
- `probeDockerCliVersion()` derives the in-container binary from the registry
rather than assuming it equals the mode name (`antigravity` runs `agy`).
- The remote CLI version probe now covers grok and deepseek, which the
hardcoded map it replaces omitted while its own comment said the rule was
"every mode except shell".
- `codeman doctor`'s CLI rows are generated from the entries, so Claude's
install hint is the install command rather than a docs URL, five CLIs gain
hints they never had, and the row order follows the catalog.
Also hardened along the way: `sessionModeSchema()` is bounded at 24 chars
(matching the `cliId` pattern) before its failure message quotes the value
back, and `deepMerge` skips `__proto__`/`constructor`/`prototype` when reading
the hand-editable `clis.json`.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WQkoi1cNegqVwZHgzx5SbJ
The header chip silently shrank from "5h 4% · 7d 52%" to a lone "7d 52%", which
reads as half the feature breaking rather than as an idle window.
Nothing was broken. Claude Code documents `rate_limits.five_hour` as "present
only while the API reports it and its resets_at has not passed", so between
5-hour session windows the key simply leaves the statusline payload. Codeman's
snapshot replaces the Claude half wholesale on every sample, so the segment
disappeared until usage opened a new window. Confirmed against a live 2.1.252
session by capturing real statusline payloads on an isolated tmux socket: the
boot render carries no `rate_limits` at all, and the first post-response render
carries both windows.
The slot now stays, with a dimmed em dash. Claude only: a missing CODEX bucket
means that plan has no such limit rather than an idle window, so those stay
omitted (pinned by the existing test). The placeholder can never stand alone
either — hasWindows() still gates the row, so a provider reporting nothing
renders nothing rather than a row of dashes. The tooltip says "5-hour limit: no
active session window" instead of dropping the line.
Verified in a real browser against a dev instance: the idle chip renders
"5h — · 7d 52%" with the dash at opacity 0.55 in --text-dim while the live
value keeps its green, and the chip holds its shape (100px idle vs 107px with
both windows).
Docker Compose deployment: Codeman runs in a container and spawns Docker cases
as SIBLING containers through the mounted host socket (Docker-outside-of-Docker).
Resolved the README conflict (master had grown to eight CLIs since the branch
was cut) and moved the Compose blurb out of the feature bullets into Quick
Start, next to the other ways of starting Codeman.
Three review findings from the PR discussion are fixed here rather than left
for a follow-up, because two of them are shipped-image problems:
- `.dockerignore` excluded `.env` only at the ROOT. A pattern is matched against
the whole context-relative path, so `docker/.env` — which the deployment's own
README tells the user to fill with CODEMAN_PASSWORD and provider API keys —
was picked up by `COPY . .` and baked into the image at
/opt/codeman/docker/.env. Verified in both directions against a real build
context: with a canary secret in docker/.env, the unfixed ignore file lets
/ctx/docker/.env through, and `**/.env` (plus `**/.env.*` and a negation for
the checked-in .env.example) leaves only the example behind.
- `CODEMAN_CASES_PATH` moved the server's CASES_DIR but not the CLI's, which
still hardcoded ~/codeman-cases, so `codeman skill install --case <name>`
reported "Case not found" on exactly the deployment the override exists for.
Both now resolve through config/cases-dir.ts. state-store.ts keeps its own
literal on purpose: that one migrates the historical ~/claudeman-cases
directory by name and is about the old default, not the active location.
- CLAUDE.md gained the Compose paragraph (the sibling-container inversion, the
three env vars, the .dockerignore and root-owned-bind traps) and .dockerignore
joins the documented list of files that genuinely belong in the repo root.
The PR's `mode === 'claude'` guard on dockerResumeId is an unrelated master bug
fix riding along: appendResumeFlag() maps a resume id onto codex/gemini/pi/grok/
deepseek/omp/antigravity and RESUME_ID_SAFE accepts a UUID, so a Docker case's
lastClaudeSessionId was handed to every non-claude CLI.
Full gate green in a merge worktree: 6360 tests, lint, format, frontend syntax,
public assets, lockfile.
Which conversation a pane is on was re-derived by correlating
~/.claude/history.jsonl against Session.lastSubmitAt — and lastSubmitAt is
bumped only by input that flows through Codeman's own write path
(Session.write / writeViaMux). A user who attaches to the pane's tmux session
directly never set it, so resolveActiveClaudeSessionIdFromHistory() returned at
its first line for that pane's whole life and the response viewer stayed pinned
to the launch conversation, showing a pre-/clear transcript indefinitely.
A UserPromptSubmit hook reports the live conversation id from inside the CLI
process, delivered under the pane's own $CODEMAN_SESSION_ID. That binding is a
fact rather than a correlation: it never consults workingDir, so it cannot be
claimed by a sibling pane on the same folder, a closed tab, or a bare `claude`
in the user's terminal. A pane holding such an id skips the correlation
entirely, so the number of prompts eligible for cwd-based guessing goes DOWN,
never up — the naive alternative (relax the guard, or synthesize an anchor from
PTY activity) is the reverted bug the resolver's own comment describes.
The hook also stamps lastSubmitAt, so it finally means "a prompt was submitted"
rather than "typed into Codeman's web terminal". Conversations vouched for
first-hand — and only those — extend a persisted claudeSessionChain, whose tail
re-pins the conversation when a surviving tmux session is re-attached after a
restart. ⚠️ start() resets the id at THREE points and the last one runs
unconditionally after the mux branch, so the tail is applied there too; patching
only the mux branch looks right and silently does nothing.
⚠️ The hook's stdout is discarded with curl's own -o /dev/null. Claude Code
injects a UserPromptSubmit hook's stdout into the model's context ("Exit code 0
- stdout shown to Claude"), and a trailing >/dev/null does NOT work: curlCmd
already ends `... 2>/dev/null || true`, and in `pipeline || true >/dev/null` the
shell binds the redirection to `true`, which never runs on the success path. The
discard is opt-in so the five SSE-fed events keep byte-identical command text
and no workspace's settings file is rewritten for them. The staleness marker is
quote-free for the matching reason: hooksJson is JSON.stringify'd, so a quoted
needle never matches and the gate would rewrite every workspace on every spawn.
Existing workspaces heal on their next Claude spawn through the staleness sweep.
The Claude reader concatenated every assistant row between two human prompts
into one card, fusing up to 74 distinct model messages into a single card, and
it never read the attachment rows that hold a prompt typed while the agent was
working. Measured over 57 real transcripts on 2026-09-01, the viewer shows
1,806 messages instead of 356 and 353 user cards instead of 178, with the
assistant text sequence unchanged row for row and the response without
?context=full byte-identical on all 57 files.
One assistant row IS one whole model message: in that corpus no assistant row
carries more than one content block and no message id carries more than one
text block, so there was nothing to reassemble. Each row becomes its own
message carrying an additive {kind, label, turn}, and the frontend renders a
same-role run inside one turn as badge-less continuation segments — which is
what keeps a p90 of 11 messages per turn from reading as card spam. A numeric
turn gates that rendering, so Codex, the external-CLI pane parser and an older
server keep one badge per card.
A prompt typed while Claude is working is recorded ONLY as an
attachment/queued_command row. Taking it when origin.kind is 'human' and
commandMode is 'prompt' recovers 162 user cards from 163 such rows — one is a
verbatim repeat inside an unanswered user run and is collapsed by the existing
dedup guard — and restores the turn boundary whose absence let the assistant
runs fuse. The CLI's own queue entries are cleanly separable: of 322
queued_command rows, 159 are commandMode 'task-notification' and not one of
them carries an origin key.
This narrows #169 rather than reverting it: sidechain exclusion, the
restored-<uuid8> rebind, replayed-snapshot dedup and synthetic-row filtering
are all unchanged and still asserted.
mobile.css hides #createCaseModal's .set-foot below 860px, and that modal's
header — unlike Settings' — carries no set-head-save. So on a phone the
Create/Link button existed nowhere and the modal could not be submitted at all.
Adds the header button and drives both together through switchCaseModalTab()
and submitCaseModal(), so whichever one is pressed the other shows the same
pending state and is equally unclickable. Following the Settings pattern also
means Add Case picks up the existing .set-head-actions:has(.set-head-save) tray
and .set-head-save sizing with no new CSS; the mobile.css comment that still
listed Add Case as a lone-× sheet is corrected to match.
Audit of CLAUDE.md against the tree. Verified still accurate: the 31-module
frontend load order (matches index.html exactly), SSE registry parity at
157 = 157 (confirmed by running the parity test), config/ 21 files, types/ 22
domain files, 136 mobile device profiles, the version line, and every Quick
Reference command.
Drift corrected: 24 route modules to 25, ~220 handlers to ~227, system-routes
51 to 56, app.js ~5K lines to ~6.7K and 30 modules to 31, install.sh 92KB to
104KB. Completed the CLI resolver inventory, which was missing
deepseek-cli-resolver and omp-cli-resolver even though both modes are
documented, and named the shared cli-executable-resolver lookup chain.
Filled the gaps found by sweeping every src module against the file:
- Owner tab layouts (COD-359) had 6 source modules, 7 test files, 2 routes, an
SSE event and a state.json key, with zero mentions anywhere in CLAUDE.md or
docs/. The paragraph records the four things a reader would otherwise get
wrong: it is backend-only as of 1.24.1 with no frontend consumer, the service
is the sole mutation boundary, it projects onto PUT /api/session-order rather
than replacing it, and reconciliation is gated on a successful restore.
- codeman doctor and codeman users were undocumented top-level CLI commands.
- Four subsystems whose invariants lived only in their @fileoverview:
the workspace-trust dialog recognizer, proc-tree's bounded walk (the
2026-07-30 incident that took a machine down), deepseek-web-server (one
child process, deliberately not a shell session), and the Files panel
search matcher (globs are never compiled to a RegExp).
Also fixes a stale "156 event types" comment in constants.js (actual: 157) and
a contradiction in AGENTS.md, which still carried the retired "never run the
full suite inside a managed tmux session" rule against CLAUDE.md's current
"npm test is the gate and is safe to run bare".
Note: the trust-dialog paragraph documents trustDialogNextKey(), which is part
of a sibling session's in-flight fix for the Claude Code 2.1.252 layout change
(unnumbered, reversed options with "No, exit" highlighted, so a blind carriage
return picks exit and kills the pane). That fix was uncommitted in the shared
tree when this landed, so the doc leads the code until it is committed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NMN8UuvdBim3iM87reuQ9Z
Follow-up to #350, which fixed the actual blocker (issue #352): `dsh plugin` is
a thin forwarder that `spawnSync`s a literal `pnpm` with no npm fallback, so an
image without pnpm dies at exit 127 and takes the whole build with it.
That PR also pinned an allowlist of the two packages whose lifecycle scripts
pnpm blocked at the time. Replace it with a policy that cannot go stale: pnpm,
unlike npm, refuses dependency build scripts by default and FAILS the install
over it (`ERR_PNPM_IGNORED_BUILDS`, exit 1, measured on pnpm 11.24), and the
names to allow move between rebuilds because `@deepseek-harness-tui/dsh-tui` is
resolved by dist-tag, not pinned: 0.9.3 pulled `@google/genai` (whose script is
a literal `preinstall: no-op`), 0.10.0-beta.x does not. An allowlist of two
names would have let the next tree break the build the same way. Allowing them
wholesale is also the exposure this image already accepts three layers up,
where `npm install -g` runs the install scripts of every transitive dep of the
five CLIs above with no gate at all.
Also correct a comment in the `/api/deepseek/install-profile` route that
asserted the opposite of what #352 proved ("dsh bundles its own package
manager, so no system pnpm is required"). The route's behavior is already
right: dsh's own "pnpm not found on PATH" stderr reaches the caller as the
OPERATION_FAILED detail, so the UI's "add a terminal profile" button names the
fix. Documented the prerequisite in docs/deepseek-integration.md, and taught
the docker-cases image smoke test about `dsh`/`omp` plus the profile check that
`dsh --version` does NOT cover.
Format check failed twice, on different files each time, because three prettier
versions were in play: package.json says ^3.4.0, package-lock pins 3.8.3 (CI runs
npm ci, so that is the one CI uses), and the local node_modules had 3.9.6. Files
formatted with 3.9.6 were then "fixed" with 3.4.2, pushing session-routes and
system-routes onto a third style — every version change moved the failure to a
different set of files.
Line-break placement in `await import` and a union type only; no logic changes.
The run menu still offered every mode for an attached container. The browser's
actual request showed why:
POST /api/docker-cases/adopt-preflight -> 400
{"error":"Invalid input: expected object, received string"}
_api serializes `body` and sets Content-Type itself, and three call sites each
passed an already-stringified body, so it was encoded twice and the server saw a
JSON string where it expects an object. curl was fine throughout, so nothing in
the server logs pointed at it.
Also fixes the design defect underneath: a failed probe fell through to "do not
gate", which silently offered every mode. When the container has been recreated,
is stopped, or the engine is unreachable, the user sees claude, clicks it, and
it can only fail — with the reason visible nowhere. A failed probe now hides
every agent mode (Shell needs no CLI and stays) and shows the server's own
reason at the top of the menu.
Two static guards switched from a character window to brace matching. They
sliced between two call sites, and _loadRunModeHistory's call appears above its
definition, so the slice came out empty and the assertion verified nothing —
the same trap twice in one file.
Both paths in the adoption form had to be typed. Each gets a Browse button
using the same path-input-group markup Link Existing uses, so the two look and
behave alike.
What they can browse differs, and that is the point. The host workspace path
reuses the existing host picker. The container workdir cannot: an adopted
container has nothing mounted at a matching host path, so a host listing would
be a different filesystem — and getting this field wrong is the source of the
opaque OCI chdir error at launch, which makes it the field that most needs to
be clickable.
Adds a read-only POST /api/docker-cases/browse: one `ls` through docker exec, no
writes, no lifecycle, path shell-escaped like every other value. `ls -Ap` marks
directories with a trailing slash and keeps names with spaces intact.
PathPicker takes an optional fetchListing source rather than being forked: the
container variant only swaps where the rows come from, and reuses the rendering,
navigation, Up and Choose/Select unchanged.
Link Existing's Browse did nothing: GET /api/filesystem/browse answered 403
"No filesystem browse roots are available".
Two rules were fighting. /root is a default blocked tree in the attachment
guard, and Codeman running as root — containers, plenty of servers — makes
homedir() exactly /root, so the picker's own allowlisted Home root was blocked;
the other candidates live under it or do not exist. The root list came out
empty and there was nothing the user could open.
The blocked trees exist to keep ~/.ssh and friends out of reach, not to seal off
the user's own home. Only trees that would swallow a configured root whole are
dropped now: /root goes when Home is it (or sits inside it), /etc holds no
configured root and is untouched. Secrets stay protected — isSensitivePath
independently matches .ssh/, .env and credentials* at any depth, and it is what
the directory probe asks about.
⚠️ Navigation must reuse the same narrowed list the roots were chosen with.
Handing the raw trees downstream admits a root and then refuses every path
inside it, which reads as a picker that opens and does nothing.
Attaching a container, picking claude and hitting Run gave one line —
`execvp(3) failed.: No such file or directory` — and the run-mode menu offered
every mode. Three separate defects, found on a real deployment.
TmuxManager.createSession resolved the CLI directory without distinguishing a
docker session, so a host with no claude threw, the catch fell back to a direct
PTY, and that PTY exec'd the CLI on the HOST. The failure surfaced as a bare
execvp error naming nothing. A docker session runs its CLI inside the container;
the host does not need it. All eight modes now sit behind a cliRunsInContainer
guard, and whether the container has the CLI is settled by the adoption
preflight or the image gate before launch.
The running check used a bare double quote and command substitution. The whole
chain is embedded in an outer `bash -c "…"`, so the unescaped quote closed that
string early and the remainder was re-tokenized. It is now a `grep -qx` pipeline
using only the single-quote form every other line in the builder already uses.
Claude Code refuses --dangerously-skip-permissions as root. Our base image runs
a non-root user, so an owned container never hit this; an adopted container's
user belongs to its owner and is frequently root, and keeping the flag killed
the pane with a message visible only inside the container. The preflight now
reports runsAsRoot and the launch chain drops the flag for it.
The menu also showed every mode because the container CLI probe only started
when the menu opened. It is warmed when the case is selected instead.
Storing the container's CLIs on the case at attach time left two gaps: a case
linked before that field existed has none at all, and a container's CLIs can be
installed or removed long after it was linked. A real deployment hit the first
one — the host had only codex, the container only claude, and with no stored
list the menu still gated on the host and hid the mode that actually worked.
The probe now runs when a container case is selected, reusing the existing
adopt-preflight endpoint, so there is no new backend surface. Results are cached
per case for the page's lifetime, since the menu opens often and the probe is a
`docker exec` round trip; a concurrent probe for the same case is deduplicated
with an in-flight marker.
A failed probe leaves the cache empty, which the caller reads as "unknown" and
therefore does not gate. Hiding every mode because one probe failed is worse
than offering one that turns out to be missing, which the launch path already
refuses with a specific message.
The repaint only happens while the menu is still open, so a late answer cannot
make the list jump under a user who already closed it.
The run-mode dropdown hides CLIs that are not installed on the HOST (#201). That
is right for local sessions and wrong for a container case, whose agents run
inside the container: a host with no claude installed hides the mode while the
container ships one, which is exactly what happened on a real deployment.
The adoption preflight already probes what the container has, so that result is
persisted on the case and surfaced through CaseInfo. Docker cases gate on it;
every other case keeps the host probe unchanged.
An absent list reads as "do not gate" rather than "nothing available": an owned
container runs our base image, which ships every CLI, and treating unknown as
empty would leave the menu with Shell alone.
Typing a container name from memory is error-prone. The field becomes a native
datalist: pick from the engine's containers, type to filter, or type a name that
is not listed (the engine may be remote, or the container may not exist yet).
A datalist gives all three natively, so no dropdown state machine is introduced.
Adds listDockerContainers and GET /api/docker-hosts/:hostId/containers, following
the listRemoteCodemanSessions discovery precedent: read-only and never throwing,
so an unreachable daemon returns an empty list and the field degrades to plain
text instead of erroring.
Stopped containers stay in the list, sorted after running ones and labelled.
Attaching does require a running container, but hiding stopped ones turns "my
container is not in the list" into a dead end, while showing
`Exited (137) 8 days ago` says exactly what to fix.
The new strings were English only. Adding entries surfaced a deeper problem: the
translator matches whole text nodes and skips `code`/`pre`, so an inline `<code>`
mid-sentence splits a hint into fragments that can never match an entry — which is
why the panel's existing "Build it once with <code>...</code>" hint was never
translated either.
Drops the inline markup from the new hints so each is a single text node, then
adds the zh-CN entries. The brand name goes through the existing {name}
placeholder.
Server-side error bodies are deliberately not added: the client receives them
already interpolated with a concrete container name, so a template key could
never match.
Attaching lived only on the Docker tab, but the place users look for anything
container-shaped is the "Run in an isolated Docker container" checkbox on Create
New. A feature nobody can find is a feature nobody has.
Adds a one-click link there that switches to the Docker tab, turns the toggle on
and focuses the container field. Reuses switchCaseModalTab and the existing sync
helper; no new CSS.
Two defects that only a real container exposes.
The probe chained `command -v X && echo X` with semicolons, and a script's exit
status is its last command's. A container without the last probed CLI made the
whole `sh -lc` exit 1, so a perfectly healthy container with tmux and claude was
reported as "could not exec into the container". A missing CLI is data here, not
failure, so the script now ends with `exit 0`.
containerWorkdir defaulted to hostWorkspacePath. That default holds for an owned
container only because the create-time bind mount puts the host directory at that
exact path; attaching mounts nothing, so the two are independent facts. A host
path absent inside the container makes `docker exec --workdir` fail with an OCI
chdir error that surfaces in the pane as a bare "execvp failed". The preflight now
proves the directory exists inside the container and refuses at link time.
The Docker tab gains an "Attach to an existing container" toggle. Ticking it
swaps the create-time fields (image, network, advanced) — which describe a
`docker create` attaching never runs — for the container name, and routes the
submit to the adopt endpoint.
Reuses the existing linkDockerCase flow end to end: only the final call differs.
The docker-host upsert still applies, since it is what resolves the
engine/context/daemon for `docker exec`; its create-time fields are simply never
read for an attached case.
Docker cases could only run in a container Codeman created itself. Attaching to
one the user already built and runs means Codeman must leave that container's
lifecycle completely alone, which the launch chain could not do: it was
`image inspect` -> `inspect || create` -> `start` -> `exec`.
Adds `DockerCase.owned`, mirroring the `owned:false` contract remote-SSH already
uses for attached sessions. Absent (every existing case) means owned, so current
behaviour is byte-identical. `false` means the container belongs to the user and
Codeman may only exec into it.
The launch chain for an attached container only looks, then execs: no image gate
(the image is theirs), no create, and no `start` — starting a container we do not
own is the very mutation attaching promises not to perform. A missing or stopped
container fails closed with an actionable message instead. Credential seeding is
skipped too: those copies read from create-time read-only mounts that do not
exist here, and writing host credentials into someone's container is not ours to
do, so its CLIs must already be authenticated inside it.
Four fail-closed guards. buildDockerStopCommand and buildDockerRemoveCommand
throw during pure string construction, so no caller bug can turn into a
`docker stop`/`rm` on a container we do not own; removeDockerContainer refuses
again at the lowest layer; drift reports "none" for an attached container, which
carries no `codeman.confighash` label and would otherwise always look drifted and
409 the launch gate forever; and the orphan reaper skips attached containers
through a check deliberately independent of the two conditions already covering
them.
`owned` is applied AFTER the config hash is computed. dockerConfigHash takes an
explicit field list, so ownership can never shift an existing case's hash — if it
did, every pre-existing case would trip the drift gate at once, and the remedy
the UI offers is "recreate the container".
Adds POST /api/cases/docker-adopt and a read-only
POST /api/docker-cases/adopt-preflight. The preflight refuses at LINK time rather
than at session launch, where the only ways out would be a dead pane or starting
a container we do not own.
Tests assert the negative guarantee directly — that create, start, stop, rm,
restart and kill are absent from the generated commands while `docker exec -it`
and `new-session -A` remain — since it cannot be observed by using the feature.
Small cleanup items from upstream review (Ark0N/Codeman#353):
- OMP_SEARCH_DIRS now leads with ~/.local/bin, matching omp.sh's real
installer target (~/.omp/bin was an earlier unverified guess, confirmed
wrong against a real --no-cache Docker build).
- docs/omp-integration.md: fixed the dead GitHub URL (can1357/omp ->
can1357/oh-my-pi), corrected the CLI count (ninth backend, tenth
SessionMode incl. shell -- not eighth), matched the install-path guidance
to the resolver fix, updated the version example to the actually-tested
18.0.8, and added a Docker-section caveat: --resume pinning does not
currently reach an in-container omp process, since Docker panes never see
ompConfig.
- docs/architecture-invariants.md: fixed a heading missing ", OMP" (CLAUDE.md
already linked to the -omp anchor, so the link was dead) and added an OMP
specifics paragraph -- the one external CLI missing an entry in this doc.
- .changeset/omp-backend.md: corrected the sibling-CLI list (was missing Pi,
Grok, and DeepSeek Harness) and the backend count.
- Removed a stray orphaned comment fragment in the quick-start docker branch
and split two CSS lines that had two declarations jammed onto one line.
resumeHistorySession() creates the resumed row in its own mode via a
modeConfigKey map (opencode/pi/grok/omp -> continueSession, deepseek ->
resumeSession) and retires the old row afterward. codex, gemini and
antigravity were missing from that map, so resuming one of their rows
started a brand-new session with NO continuation while still deleting
the row it came from -- silent data loss dressed as the duplicate-row
fix. Gate row retirement on continuesSomething (true only for modes that
actually got a continuation config) instead of wiring an unverified
sessionId->native-conversation-id assumption for the three affected CLIs.
DELETE /api/sessions/:id reimplemented the ownership 404 check inline in
two places instead of going through findSessionOrFail, and its
persisted-only-session branch never broadcast session:deleted, so other
open tabs kept the retired row until their next unrelated fetch. Extract
the shared 404 into sessionNotFoundError(), add findPersistedSessionOrFail()
alongside findSessionOrFail() in route-helpers.ts (same ownership
contract, returns a SessionState instead of a live Session), and use both
from the route instead of inline checks. Add the missing broadcast.
The docs claimed omp "has no documented vendor-key namespace of its own"
and "the multi-user clamp has nothing to gate" for omp — both false. Per
omp's own docs/environment-variables.md, it reads ~40 provider keys from
env (pi's known 34-key problem in the same shape), and its own knobs are
mostly PI_* (already globally allowlisted): PI_CONFIG_DIR,
PI_CODING_AGENT_DIR, PI_CODING_AGENT_SESSION_DIR, PI_SUBPROCESS_CMD,
PI_SHELL_PREFIX. The first three also move the ~/.omp tree
omp-session-resolver.ts/omp-transcript.ts hardcode, silently degrading
pinning/history — a known gap shared with pi, documented but not fixed
here.
The OMP_* prefix this PR adds brings in OMP_AUTH_BROKER_URL/
OMP_AUTH_BROKER_TOKEN, where omp resolves credentials from — the same
shape DEEPSEEK_BASE_URL is already dropped for in
clampEnvOverridesForOwner(). Add both to OWNER_CLAMPED_ENV_KEYS so a
non-granted owner in multi-user mode can't redirect them, and correct the
false claims in CLAUDE.md, docs/omp-integration.md, and the stale
resolveOmpHome() comment. Also documents omp's default
tools.approvalMode: yolo, which was previously unstated.
Follow-up from a full-branch review pass (Opus) of the omp-mode integration:
- Add pinning tests for resolveOmpConfigForCreate() (session-routes.ts),
exported to make it testable: the exact "resume this OMP row from
history" pipeline that mangleOmpWorkingDir's earlier bug lived in had
zero coverage despite being the resolver module's whole reason to exist.
- Log a warning when findLatestOmpSessionId() finds nothing on disk and
continuation silently degrades to omp's own ambiguous --continue,
in both call sites (session create and respawn pinning) - previously
silent, making the degradation invisible to anyone debugging it.
- Require an absolute cwd before trusting a session file's working
directory in omp-transcript.ts's parser, so a corrupted/malformed
session file can't point a downstream resume at a relative or empty
path.
- Document (don't speculatively fix) an unverified symlinked-$HOME edge
case in mangleOmpWorkingDir(): the review's suggested realpath() fix
assumes omp itself resolves symlinks before mangling, which is
unconfirmed - guessing wrong there would trade one silent mismatch
for a different one.
- Incidental: fixed unrelated pre-existing prettier drift in
session-routes.ts (antigravity/opencode dynamic import line-wrapping)
that was blocking the pre-commit formatting gate on this file.
Confirmed as a non-issue: the model-name regex allowing "/" is
intentional (provider/model ids like "crof/glm-5.2" were used
successfully in live testing).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude conversations survive "Kill Tmux & Claude" because Codeman reads
them back independently from ~/.claude/projects, not from its own
session bookkeeping. omp conversations had no equivalent: kill the
Codeman session and the conversation vanished from Past Sessions
entirely, even though omp itself never forgot it on disk.
Adds omp-transcript.ts, a scanner over omp's own
~/.omp/agent/sessions/<mangled-cwd>/<uuid>.jsonl files (the same shape
as Claude Code's own transcript scanner, but simpler -- these files are
small enough to read whole instead of doing head/tail windows). Each
file's own "session" header line carries the real cwd and session id
directly, so unlike Claude's mangled-directory-name decoding this
never has to guess. Wired into gatherUnifiedInputs() as a second
history source alongside the Claude scan, and HistoryInput/
mergeUnifiedSessions() now carry an optional `mode` so a non-claude
history-only row still gets a real mode badge.
Also fixes the ambiguity behind the "continue picks the wrong
conversation" report from this session's testing: omp mints its OWN
session uuid, unrelated to Codeman's, so a live/persisted row and its
own history-scan row would otherwise show up as two separate entries
for the same conversation the moment the id gets resolved. Reuses the
existing claudeSessionId alias field (mergeUnifiedSessions' fold-into-
owner mechanism) to point at the resolved omp id, threading it through
every place `_claudeSessionId` gets (re)computed -- the constructor,
_resolvedOmpRespawnConfig, and a new _maybeCaptureOmpSessionId() that
opportunistically resolves it the first time a brand-new omp session
(one that has never gone through a respawn) goes idle.
Also closes a THIRD instance of the "ompConfig never got wired in
here" gap this session kept finding: restoreMuxSessions() in server.ts
restores every sibling CLI's config from persisted state on boot except
omp's, so a boot-recovered omp session always lost its resolved resume
id and fell back to guessing again.
Verified live end-to-end: told a session a secret, killed it fully
(Kill Tmux equivalent, killMux=true -- the Codeman session AND its tmux
pane both gone), and the conversation still showed up in the unified
list as a history-sourced row with the real first prompt as its title
and an omp mode badge, keyed by omp's own session id.
Known remaining gap, not fixed here: the claudeSessionId alias doesn't
yet resolve reliably on every boot-recovery path for a session that
was never respawned while alive (e.g. a plain re-attach to a pane that
was never dead) -- worth a follow-up, but doesn't affect the two things
that matter most: the conversation surviving a kill, and continuation
correctness once an id has been resolved (which happens on the very
next respawn either way).
Every non-claude "Resume" click creates a brand-new Codeman session
(there is no id to reattach to), but the old row was never cleaned up
-- click resume on the same conversation a few times and the session
list fills up with duplicate rows sharing one name. resumeHistorySession
now retires the row it resumed from after the new one starts.
That retirement needs DELETE to actually work on a row that was never
live in the first place (the normal case for anything showing up in
"Resume Conversation"): findSessionOrFail only checks the in-memory
live-session map, so DELETE 404s on a persisted-only entry today. Give
the route a fallback: when the id isn't live, look it up in persisted
state instead and demote/remove it there (respecting the existing
pinned-session protection). Verified live against a real persisted-only
row via the API, and added route-test coverage for both the success
and still-truly-unknown-id cases (which needed a demoteOrRemoveSession
mock the route harness didn't have).
Also includes an unrelated pre-existing prettier drift fix picked up
by npm run format (omp-cli-resolver.ts, antigravity/opencode import
wrapping in session-routes.ts).
resumeHistorySession() never sent mode when recreating a session from a
history/session-manager row, so the server default silently opened a
plain Claude session for every non-claude row -- reproduced live: OMP
rows spawned Claude sessions on click. Thread the row's mode through
every call site (welcome list, session manager, mobile overview) and
only send the Claude-specific resumeSessionId for claude rows.
Codeman has no live PTY-reattach outside server boot, and it's moot for
OMP anyway (exiting it kills the pane's only process), so route the
non-claude relaunch through each CLI's own continue-most-recent flag
instead of a context-free fresh start. OMP never got one: buildOmpCommand
only implemented --model/--resume despite omp --help documenting
-c/--continue. Added continueSession to OmpConfig end-to-end (type,
schema, builder) mirroring the existing opencode/pi/grok/deepseek
fields, and wired resumeHistorySession to use it.
Verified live: told a real omp session a secret, exited it, closed the
tab without killing tmux, relaunched with --continue in the same
directory, and had it recall the secret.
The chip prefixes every row with the provider name, so a machine that only
has Claude limits renders "CLAUDE 5H 60% 7D 23%" — a 46px label naming the
only thing it could possibly be. The name exists to tell two rows apart, so
it should only appear when there are two.
updatePlanUsageChip() now checks whether both Claude and Codex actually have
windows before building the rows, and emits the .pu-provider span only in
that case. The tooltip keeps naming the provider in both cases: it has the
room, and the chip no longer does.
Verified in a browser on an isolated beta instance: Claude-only renders bare
windows with no .pu-provider in the DOM, Codex-only the same, and the
two-provider chip is byte-identical to before.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The rebase hand-repair dropped the closing brace of .welcome-btn-pi:hover
and .btn-toolbar.btn-run.mode-pi:hover before the inserted OMP rules.
The browser CSS parser drops every rule after an unclosed block, so the
deployed UI rendered as unstyled text bars (only ~456 of ~2583 rules
applied). Verified clean via esbuild --minify (no css-syntax-error) and
rebuilt dist.
install.sh installs a build toolchain on Linux (node-pty has no Linux
prebuild, so a stock Ubuntu 24 server died inside node-gyp with
"not found: make"), plus review hardening for #339: the write-queue
reset paths now release the one-chunk-in-flight gate.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Fifteen review findings on the dsh mode, the serious ones first:
- Multi-user: DEEPSEEK_BASE_URL joins the owner-clamped env keys.
_configureDeepSeek() forwards the SERVER's own DEEPSEEK_API_KEY into
every dsh pane and applyEnvOverrides() lands after it, so a non-granted
owner who could redirect the base URL would have the operator's key sent
as a bearer credential to a host of their choosing.
- Wait registry: until=stop/blocked is refused on docker and remote-SSH
dsh sessions (new deepSeekBridgeUnreachable fact in sessionHookOptions).
The HERDR triple is set via LOCAL tmux setenv, which crosses neither
docker exec nor ssh, so such a session can never post a hook event and
the wait burned its whole timeout on every turn.
- Approvals: a dsh item is an ALERT, not an answerable card. The answer
route refuses (the '1'/Esc keystrokes are Claude-dialog-shaped and the
option parser cannot read a third-party TUI's frames, so an answer was a
blind keystroke into a foreign composer), and the push notification
carries no Approve/Deny actions for dsh sessions.
- Status shim (v3): --seq is forwarded and the server drops stale retried
reports inside a 60s window (the TUI retries with backoff, so a retried
'working' could land after 'blocked' and resolve an approval whose
dialog was still on screen); 4xx responses exit 0 instead of retrying,
so one misconfigured session cannot feed the auth rate-limit bucket
until the hook endpoint 429s for the whole instance.
- Web-UI server: concurrent starts are serialized through a lock (two
racing POSTs used to pick the same port and orphan the winner), and the
readiness poll / timeout paths only clear or stop the singleton while it
is still theirs. First click actually opens the tab now
(refreshWebviews, not the nonexistent loadWebviews). DELETE
/api/deepseek/web requires the privileged grant in multi-user mode.
- Cron: deepseek jobs run the same two-part launch gate as the HTTP
create paths (impl moved into the resolver so all three share it) and no
longer stamp a Claude default model on the session.
- Parity sweeps: quick-start's docker branch rejects deepSeekConfig like
the remote branch; the Ralph auto-enable list gained deepseek;
HookEventType gained agent_working; the phone overview run menu filters
managed webview records like the desktop menu.
- install.sh: the dsh identity probe closes stdin (under curl|bash a
child that reads stdin eats the rest of the script), bounds the exec
with timeout where available, and is memoized to one scan per install.
- Welcome screen: .welcome-btn-deepseek styled in the #4d6bfe brand
identity (it rendered as an unstyled UA-grey button); stale markup
comment about the web shortcut rewritten; clamp docs updated.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Four review findings on the worker-transcript feature:
- Docker and remote-SSH dsh sessions now keep the pane segmenter: their
transcripts live in the container's / remote host's own ~/.dsh, which
the local reader can never see, so the transcript path returned
'nothing said yet' forever and an agent polling such a worker starved
on an answer that existed. Gated on !session.docker && !session.remote
(statically pinned) and documented in the integration guide.
- last-response reads are memoized on (path, mtime, size, blocks): the
skill's last_text polls once per second, and each poll decompressed and
reparsed the whole file on the event loop even when nothing had been
appended. An unchanged poll now costs one stat.
- The pairing ladder's comment claimed /new is served by step 2; in truth
the boot-window transcript wins for as long as it exists (deliberately:
preferring newest-eligible would hand a worker its busier sibling's
reply). The comment now states the real tradeoff instead of the
aspirational one. Same for decodeZstdFrames' 'skipped' wording — a
corrupt frame truncates the decode there, which is the safe behavior.
- stripReasoningPrefix no longer runs on user prompt text, so a prompt
containing a literal </think> renders whole in blocks view.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Three review findings on the detailed-rows feature, all in its edge cases:
- The App Settings width select consulted the handheld defaults blob
(tabRailWidth: 256) BEFORE the rich-aware default, which the renderer
never reads — so a tablet's unsized rich rail rendered 320 while the
dialog said 256, and a routine Save persisted the 256 (below the 288px
tight threshold, permanently). The chain now mirrors
applyTabRailWidth()'s actual resolution.
- _setTabRailWidth() re-rendered on a compact flip but never re-ran
applyTabWrapSettings(), the one owner of the folder line, whose railRich
input reads the compact class this function just toggled. A rich rail
dragged below 240px kept emitting folder rows — persistently, for a
stored width < 240, since the boot wrap pass runs before the class is
first applied. The wrap pass now re-runs on the flip, with exactly one
render either way.
- Both reset affordances (handle dblclick, Enter on the handle) reset to
the hardcoded 256 even on a rich rail, landing it below the tight
threshold; both now resolve the rich-aware default (320), via a new
optional defaultWidth input on resolveTabRailKeyboardWidth().
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
`GET /api/sessions/:id/last-response` is how an agent (and the Response
Viewer) reads what a worker said. DeepSeek was falling through to the
pane segmenter with the other external CLIs, which for this mode is not
merely coarse but wrong: dsh-TUI paints a full-screen splash, so a
`last-response` call on a fresh dsh session answered with its ASCII-art
logo -- and anything polling for a worker's first reply reads that as a
reply.
dsh does not belong in that group. It writes a structured JSONL
transcript per session, so read it. Four things in that file shaped the
reader, all measured against real transcripts on disk:
1. dsh appends ONE ZSTD FRAME PER WRITE, and Node's zlib zstd decoder
(one-shot and streaming alike) stops at the first frame end: a real
56-line transcript decoded as 1 line / 158 bytes -- the session header
alone, i.e. a silent truncation that reads as "nothing said yet"
forever. `zstdFrameRanges()` walks frame and block headers to find
exact boundaries; splitting on the 4-byte magic would corrupt
everything after a magic sequence occurring inside compressed data.
zstd is resolved at RUNTIME because it landed in Node 22.15 while the
project floor is 22.0, so an older Node keeps the pane behaviour.
2. Every turn also records a plugin-sourced `user/message` (the runtime
context snapshot), which must not render as the user's own words.
3. A turn that ends in an error carries the provider's message; it is
surfaced as `Turn error: …` (and a non-error early stop as
`Turn ended: …`) rather than as an empty string, which an agent reads
as "still thinking" through fifteen polls.
4. Reply text is assembled per (turn, step): a finalized message wins and
the streamed deltas fill in only for a step that never finalized, so a
partial answer is readable mid-turn and never doubled. "Finalized" is
tracked as a set of steps rather than as non-empty text, because a
step whose whole reply was reasoning strips to '' at the `</think>`
boundary and would otherwise resurrect the raw deltas in its place.
Session-to-transcript pairing is by the transcript's own header `cwd`
plus a boot window against the session's createdAt, never by
reproducing dsh's directory mangling (already two forms on disk) and
never by newest-mtime alone -- mtime alone handed a freshly spawned
worker its predecessor's answer in the same case directory.
An empty result still wins over the pane; only a Node that cannot decode
zstd falls back to it.
Clicking "DeepSeek web UI..." opened two tabs: the web tab asked for, and a
shell tab running the server next to it. The shell was deliberate - the server
lived in an ordinary session so it was visible, scrollable, killable and died
with its tab, and nothing new had to supervise a long-lived HTTP server. That
reasoning was sound and the result was still wrong in use: opening a dashboard
should open one tab, and after the first launch the terminal is pure noise.
The server moves to a background child process owned by a new
`src/deepseek-web-server.ts`, behind `POST /api/deepseek/web`. What the session
gave away for free is now explicit, which is most of the module:
- Exactly one server. A second click reuses the running one instead of racing
it for a port; the session flow could not do this at all, because two clicks
were simply two sessions.
- Restarted when the requested authority changes. `--trusted-host` fences dsh's
own /api against the browser authority, and a Codeman reachable at both
loopback and a tailnet name has two. Reusing a server fenced for the other
origin renders a page whose every call 403s, which reads as a broken
dashboard rather than a misconfigured one, so a mismatch restarts instead.
- Killed on shutdown. The child is detached so its whole plugin tree can be
signalled at once, which also means it would outlive Codeman and hold its
port against the next start - the exact EADDRINUSE this feature already got
wrong once.
- Boot output captured and returned. With no shell tab there is nowhere else
for a stack trace to land, so a failed spawn reports its own tail.
The endpoint is fenced at the same bar as the profile installer and for the
same reason: booting a dsh profile executes the plugin code in it, so this is a
privileged action even though it reads as "open a page". `authority` comes from
the client (`location.host`) because only the browser knows which origin is in
play, and it is regex-confined at the schema boundary - defence in depth behind
the argv-array spawn, admitting host:port in the shapes a browser authority can
take and nothing readable as a second argument.
`GET /api/deepseek/web-port` is gone; port selection moved into the supervisor,
which is the thing that knows whether a server is already running. The two
client-side probe helpers went with it, since the server now owns the wait.
Verified over the tailnet authority end to end: no session is created (session
count unchanged, one tab), the server runs on 3081 beside the user's own dsh
web on 3080, status reports the tailnet authority, and the proxied dashboard
renders with zero 4xx. Full gate green (6148 passed, +6).
The `Run > DeepSeek web UI...` shortcut failed three ways at once against a real
install, and the three are independent.
1. It hardcoded `--port 3080`. That is dsh web's OWN default, which makes it
precisely the port a DeepSeek user is most likely to be serving on already,
so the launch died with EADDRINUSE against the user's own server. The port
now comes from `GET /api/deepseek/web-port`, which walks 3080..3119 for a
free loopback port by BINDING it (a connect probe cannot tell "free" from
"listening but not answering yet").
2. It opened the tab unconditionally. The crashed server left a saved dashboard
pointing at nothing, with the failure only visible in a shell tab nobody had
a reason to look at. The launch now polls the existing webview probe until
the URL answers, and on timeout reports the error naming the shell tab
instead of persisting a dead dashboard.
3. The saved tab was untrusted, so the frame was sandboxed without
`allow-same-origin` and the dashboard was broken twice over: the dsh
client-runtime reads `localStorage` while loading its plugins and died there
("the document is sandboxed and lacks the 'allow-same-origin' flag"), and an
opaque-origin frame sends `Origin: null`, so dsh's own trust fence 403'd
every `/api` call no matter which authority `--trusted-host` named. Passing
`location.host` only means anything once the frame actually carries that
origin, so `--trusted-host` had never once done its job. The managed tab is
now created `trusted: true`.
That trade is real and deliberate: a trusted proxied frame is same-origin
with Codeman and can reach Codeman's API. It is defensible only because this
dashboard is an agent harness Codeman just started itself, on loopback, which
can already run code as the user. It is not a precedent for trusting
third-party dashboards, which is why it is set at this one call site rather
than defaulted.
Separately, the shortcut listed its own dashboard twice: once as the menu entry
that starts it and once as the row that entry had written on the previous click.
Webviews now carry an optional `managed` marker, managed rows are filtered out
of the saved-dashboard list, and a relaunch repoints the existing row rather
than stacking one dead dashboard per restart (which the per-launch port would
otherwise guarantee). `managed` is declared in the schema because a plain
`z.object` strips undeclared keys, so an undeclared marker would never survive
the round trip.
`DEEPSEEK_WEB_PORT` is gone from constants.js; its doc comment asserted that a
hand-started `dsh web` and the shortcut "land on the same place and share one
saved tab", which is the bug stated as a feature.
Verified on a real install with the user's own `dsh web` holding 3080: the
shortcut takes 3081, the server answers, exactly one DeepSeek entry shows in the
run menu, and the proxied dashboard renders its workspaces and completes its own
API calls (the previously-403'd `api/settings.describe` now succeeds). Full gate
green (6142 passed), typecheck/lint/format/public-assets clean.
Two review nits on the vertical rail's detailed rows.
1. The tab-rail-tight rule (below 288px) hides `.tab-meta-created`, and its
comment claimed the value "survives in the row's title attribute either way".
It did not: the only title carrying it lived ON that element, and a
`display: none` element has no hover target, so the created stamp was not
shrunk but gone with no way to ask for it. Rather than just correcting the
comment, `_sidebarRichMetaHTML()` now puts BOTH absolute stamps on the
`.tab-meta` line itself, so the pill and the gaps around the stamps remain as
hover targets. An item's own title still wins where the item is visible.
2. applyTabOrientation() decided whether applyTabWrapSettings() had already
re-rendered by comparing `_tallTabsEnabled` before and after. That reads an
UNDEFINED previous value as "it rendered", but applyTabWrapSettings()
deliberately renders nothing on its first call ever (it only establishes the
baseline: `prevTallTabs !== undefined && prevTallTabs !== showFolder`). So on
a first call that also flips the folder row, neither function rendered and the
rows stayed stale. Reachable when the pre-paint script throws and leaves the
layout attributes on their catch-branch fallbacks for applyTabOrientation() to
correct. The guard now mirrors applyTabWrapSettings()'s own condition.
Both new tests were run against the unfixed code first and fail there, which is
the only thing that makes them regression tests. (The third, "does not render
twice", passes either way by design: it pins that fix 2 did not introduce a
double rebuild.)
Verified in a real browser against a live server with two sessions, driving the
narrowing through _setTabRailWidth() the way the resize drag does: at the 320
default the row reads "CREATED 2m ago · IDLE <1m" with the created element
displayed; at 256 the tight class is on, the created element computes to
display:none, the visible text drops to "IDLE <1m", and the meta line's title
still reads "First created: ...". At 220 the compact threshold drops rich rows
entirely. Screenshots confirm no truncation artifacts in either state.
Full gate green (6104 passed), typecheck, lint, format, frontend-syntax and
public-assets all clean.
