host ran the live test against green. Two results.
SETPRIV WORKS. The privilege drop lands on the member and the SDK spawn survives
it, so the design does not change shape and everything layered on the hook
stays. That was the last question that could have moved the architecture.
THE BINARY CHECK REFUSED A BYTE-IDENTICAL PATH. sameFile used realpathSync,
which has to stat inside a 700 home the platform is `other` to, so it threw
EACCES and the catch turned that into "not their binary" — every member turn
refused forever, the moment the gates moved. Failing closed was the right
direction and it made the feature impossible rather than unsafe.
Now resolve(command) !== claudeBinIn(run.home). Both operands are computed by
the platform from the same function, so string equality establishes exactly what
the check is for and needs no access to their home. sameFile and its tests are
deleted: a helper kept for a case that cannot arise is a trap for the next
reader. The realpath version was defending against an upstream that normalises
paths, and there is no such upstream — the platform controls both ends.
The fact that decides this is that the check runs in the PLATFORM process, not
the member's, and neither of us stated it until EACCES did.
THE TEST WATCHED THE WRONG PROCESS. child.pid is sudo, whose real uid is
legitimately the service user's until it execs down through setpriv, so
asserting on it fails on a working drop. Split: --version through the hook
proves their binary ran, and a second spawn creates a file in their home whose
owner the test stats. Nothing self-reports, and a process cannot forge the uid
that owns a file it created.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
host caught a circularity I had written twice: do not lift the chat gates until
a member turn has been watched running, but a member turn goes through chat and
chat refuses non-owners. With the gates up there is nothing to watch; with them
down the thing we wanted proven has already shipped.
spawn-as-member.live.test.ts calls spawnClaudeAsMember directly against a real
provisioned account — no gate, no chat, no SDK. The child's uid is read from
/proc/<pid>/status, so it is the kernel's answer rather than anything the child
chose to say, and it asserts >=1000 and not this process's uid: a failed
privilege drop cannot pass by running as the service user. It also asserts the
binary exited 0 having printed a version, which proves their install ran rather
than merely being spawned, plus a negative that /bin/sh through the same hook
throws.
Opt-in via OFFICER_TEST_MEMBER and OFFICER_TEST_MEMBER_HOME, because it needs a
provisioned member with claude installed — which exists on the production host
and on no developer machine. A run without them skips loudly rather than
reporting an empty file as a pass.
Also adopted host's NO REPLY NEEDED terminator: "reply to everything" had no
exit condition and cost the owner two agents being polite at each other.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
host found this while we were elsewhere: four agent-sidecar crashes tonight, one
truncating the owner's turn mid-sentence.
error: ProcessTransport is not ready for writing
at write (…/claude-agent-sdk/sdk.mjs) <- no frames from our code
A floating rejection inside the SDK's own input pump, so no await of ours could
have caught it. With no handler anywhere in src/servers it reached the top
level, Bun exited, PM2 restarted, and every live session on the machine died —
not just the one whose transport failed.
That is 975673a for the second time. It fixed the one path someone had thought
of, a Postgres query throwing, and its own message named the consequence: "any
Postgres restart killed every live agent session on the machine". The general
case had no backstop.
Now logged loudly and survived. This does not pretend an unhandled rejection is
fine — it makes it debuggable instead of fatal.
uncaughtException is deliberately not handled the same way: a rejection leaves
the process's state intact, a synchronous throw that unwound to the top may not
have, and continuing on a corrupted heap is worse than restarting. Same blast
radius either way, which argues for durable sessions rather than for surviving
anything at any cost.
The trigger is untouched and unconfirmed — most likely the CLI exiting while
streamInput is still pumping. That needs catching in the act on the live host.
Also worth knowing, since d59adbf interacts: this restart sweep runs several
times an hour on that machine, and a session with no recorded userId is now
skipped rather than asked about, so it stays marked generating until a
reconnect. A stuck "generating" spinner after an agent restart is that, not the
UI.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The control surface half of 7cb402b, and the code-side blocker on the gates.
kill, interrupt, clear-session, is-generating, find-session and list all took a
bare sessionKey, so any caller who could reach them could act on whichever
session happened to match — and list returned every session in the sidecar,
which host rightly called a disclosure on its own, before anyone kills anything.
All six now carry userId, resolved from the authenticated request and never
taken from the client, and every handler enforces it through one ownedSession
helper. list is filtered rather than labelled. find-session is scoped because it
is the reattach hinge: a browser holding a transcript uuid it should not have
would otherwise be handed the session key that drives it.
"Not yours" and "does not exist" answer identically everywhere, which is the
same choice getClaudeSession made: every caller treats them the same, and a
distinct answer for the second confirms to a guesser that a session exists under
a key they do not own.
One behaviour change beyond the scoping. endTurnIfAgentIsGone sweeps sessions on
a sidecar restart, and a session with no recorded userId now has no safe id to
ask as — asking as the owner would answer a member's orphaned session with the
owner's authority. It is skipped, so it stays marked generating until the next
reconnect corrects it, which is what happened before that loop existed.
This removes the code-side reason the gates cannot move. It does not make them
movable: no member has signed in, no member turn has run, spawnClaudeCodeProcess
has still never been called, and lifting them was never mine to decide.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
host found this reading 10: chat sessions carry no identity at all. state.ts
held a flat sessionKey -> transcript uuid map, the in-memory sessions Map was
keyed the same way, and websocket.ts takes sessionKey and resumeSessionId
straight off the client message. 4d4a253f fixed exactly this for the pty
sidecar — "re-attaching to a session belonging to another account is refused,
otherwise a member resumes someone else's shell by guessing an id that travels
in a query string" — and chat never got the same treatment, because both gates
made it unreachable and therefore invisible.
Sessions now carry userId, persisted and in memory. getClaudeSession requires
the caller and returns undefined on a mismatch rather than throwing, since a
throw confirms that someone else's session exists. spawnClaudeStreaming throws
when a live session's owner does not match — that is the path that mattered
most, because handing over another account's sessionKey would otherwise push a
turn into their conversation and stream their agent's output back.
Legacy string entries are adopted to the owner on load. That is a statement
about the past rather than a guess: until this commit the gates refused every
non-owner, so nothing else could have created one. Dropping them would have
silently broken the owner's resume on upgrade.
PARTIAL, and the doc says so plainly: claude:kill, :interrupt, :clear-session,
:is-generating, :find-session and :list all still take a bare sessionKey with no
ownership check, and :list returns every session in the sidecar. Closing them is
a wide mechanical change across the protocol, the registry verbs and their
producers, and it belongs in its own reviewable commit rather than buried under
a state migration. The gates must not move on the strength of this one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
host measured what df450318 assumed. writeFileSync passes `mode` to open(2),
which honours it only when it CREATES the file — on an existing one the call
truncates and writes and the mode is ignored. So the fix worked on a fresh
install and did nothing at all on every box already leaking, which is the whole
exposed population. Verified on production after the commit: still 0644, still
readable by a member.
That is worse than not fixing it, because it closes the ticket. The exposure
would have continued through every bootstrap with nobody watching for it.
chmodSync after the write, both kept — the creation mode closes the window
between open and chmod on a fresh write, and chmodSync is what reaches an
install that is already leaking. Commented so neither is deleted as redundant.
Still not closed on disk: the file is the owner's to chmod and the token is
theirs to rotate, and no commit reaches either.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
host found a live credential exposure while answering my question about what
else the member branch missed. It was outside the diff, and predates all of it.
MCP-CONFIG WAS NOT BRANCHED. mcpHostPath is module-level, written once at the
owner's bootstrap, and was applied to every turn. Its env block carries
OFFICER_AUTH_TOKEN, a 30-day JWT signing as the owner — so a member's turn would
have spawned their MCP server holding it. Now inside the params.member ternary
alongside the binary and the spawn, for the reason already written there: these
values say whose turn this is and have to move together. A member gets none.
What they should get instead is undecided, and undefined beats the owner's.
THE FILE WAS 0644. Written with a bare writeFileSync into a 755 directory, on a
host where `terminal` is granted to every role by default — so any member could
cat it and hold owner-level API access on loopback. host verified that as a real
member on the production host rather than reasoning about it. Now 0600.
The mode is the only half of that which is code. The token has been
world-readable and stays compromised until rotated, the directory chain above it
is still 755, and neither is fixable from a commit. Both written up for the
owner in COMMS 05, along with why I am stopping here rather than continuing:
the next commit should be the rotation, not more feature work stacked on top of
an open exposure.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
ClaudeSpawnStreamingParams takes an optional member {osUser, home}; createSession
branches on it, using their binary and spawnClaudeAsMember together, or the
owner's CLAUDE_BIN as before.
THE BINARY AND THE PRIVILEGE DROP ARE ONE BRANCH ON PURPOSE. settingSources
makes ~/.claude authoritative for settings and ~ is whatever HOME the process
gets, so pointing the SDK at a member's binary while spawning as the service
user would read the OWNER'S settings and credential while running the member's
code — and it would look like it worked.
cwd defaults to member.home before HOST_HOME for the same reason: HOST_HOME is
this process's home, so a member would start in a directory they cannot read and
the failure would present as a broken agent rather than a wrong cwd.
Nothing populates `member`. Both gates refuse non-owners before any of this is
reached, so the delta is that spawnClaudeAsMember now has two importers instead
of one, and neither path a user can take changes. Verified rather than assumed,
since host made it a condition: both gates intact, 84 tests pass.
Not authorization: host gave an opinion on wire-first and deferred to the owner,
who has not ruled. Corrected in COMMS, where 01 had overstated it. The gates
come off on the owner's word alone; this reverts as one commit if the answer is
no.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
host was right twice, including about his own advice. Two dead guards had
shipped here, both for the same reason: written inside the spawn closure, where
the only way to reach them is to spawn — and the passing path spawns sudo. So
nothing ever demonstrated either one firing.
THE SUBSET CHECK WAS ALSO TAUTOLOGICAL. `permitted` came from the same constants
memberEnv builds childEnv from, so it was empty under every edit where that
holds — the exact criticism that retired NEVER_ENV. Worse, it lost the one live
trigger the denylist had: a credential added to ALLOWED_ENV used to throw, and
under the subset check widened the permitted set in the same motion and passed
silently. That is the realistic future edit and it was the one left unguarded.
Now both, and the denylist tests the LIST rather than the instance, so it fires
on exactly that edit. Extracted as `assertEnvSafe` so a test can pass a poisoned
allowlist — the guards being untestable in place is why they were decorative
twice.
THE BINARY CHECK WOULD HAVE THROWN ON EVERY TURN. Anthropic's installer puts a
symlink at ~/.local/bin/claude into a versioned directory; resolve() does not
follow symlinks, so the string compare matched only while `command` arrived as
the symlink spelling, and would have failed the moment anything upstream
normalised it — at exactly the point the hook gets wired. Compared through
realpathSync on both sides now, per turn and never cached, since `claude update`
moves the target.
Nine tests pin all of it: a poisoned allowlist, a stray key, each NEVER_ENV name,
and the symlink/target/missing-path cases.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Four fixes from the live server's review. One was a real defect.
THE BINARY WAS NEVER CHECKED. spawn-as-member passed `command` from the SDK
through untouched while a comment claimed the member's own install was what ran.
Since claude-manager resolves the OWNER'S CLAUDE_BIN at module load, wiring the
hook would have exec'd the owner's binary as the member — the precise confusion
this file exists to prevent, asserted in prose and enforced nowhere. Now throws
unless the command resolves to claudeBinIn(run.home).
NEVER_ENV COULD NOT FIRE. It tested an environment that memberEnv builds from
ALLOWED_ENV, so a denied name was already impossible; it was also missing six
credential variables the installed SDK reads. Replaced with the subset check the
reviewer proposed: anything not in ALLOWED_ENV or {HOME, CLAUDE_CONFIG_DIR} is a
leak whatever it is called. Complete by construction, and it cannot rot as the
SDK grows variables — which the denylist provably had already.
Also: one derivation of the binary path instead of two (install resolved from
the email, exec from the home — fine until they disagree), and the constraint
that ALLOWED_ENV may never hold a secret written at the list itself, since
`env K=V` in the argv is visible in /proc/<pid>/cmdline to every account.
Not acted on, and said so in COMMS: their finding that the 711 in 401dcb7 is
inert, and that a retrofit needs a mode pass. Both are theirs. Nor pulled chat
from DEFAULT_ROLE_CAPABILITIES despite agreeing a member currently sees a tile
that 403s — that is the owner's call, not a defect.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
First half of per-user Claude. Provisioning and the privilege drop, not yet
wired to a turn — the chat gates stay up and behaviour is unchanged for
everyone. Committed unfinished on purpose so the reasoning is on the record
before the server agent runs any of it; the state is written up in
COMMS/sidecar-app-store/2026-08-11-per-user-claude-handoff.md.
THE CLAIM THAT CHANGED. docs/per-user-linux-accounts.md:226-229 says the Agent
SDK "has nowhere to put a uid", so a member's turn has to become its own
process — a change of shape rather than a flag. It is a flag:
sdk.d.ts:951 exposes spawnClaudeCodeProcess, documented for exactly this ("run
Claude Code in VMs, containers, or remote environments"), and node's spawn
already satisfies the SpawnedProcess shape it wants. So no second sidecar, no
PM2 entry, no inverted transport, and none of the registry rework a second
instance would have forced (registration is name-keyed and evicts its
namesake; the nine claude verbs resolve by capability with no selector).
THE PLATFORM NEVER RUNS AS A MEMBER. The tempting reading of "each member runs
their own Claude" is a second officer-agent under their uid, and it is wrong:
that sidecar needs POSTGRES_URL and the JWT signing secret, so a member-uid
process holding them could read every account and sign a token as the owner —
strictly more than their shell can do, and already forbidden by the .env boot
check. The harness stays the service user's; the thing that runs the member's
code and holds the member's credential is theirs. That is the pty sidecar's
shape, not a new one.
PER-MEMBER BINARY, deliberately, over one shared /usr/local/bin/claude. The
private part is the credential, not the executable — but claude updates itself,
and a root-owned binary is one a member cannot update, which turns "my agent is
a version behind" into a request to the owner. Same installer the owner's own
install uses, run as them, in their home. Idempotent by skipping when present
rather than re-running: the retry button reprovisions on every press.
ALLOWLIST, NOT A FILTER, for the child's environment. At the moment of the call
the calling process holds POSTGRES_URL, the JWT secret and the owner's
ANTHROPIC_API_KEY; setpriv --reset-env means nothing crosses unless written
into the argv, so an allowlist is the complete answer to what a turn can see,
and a denylist would have to be right about every variable added later.
NEVER_ENV throws rather than leaks if someone widens it.
Login is the member's own act against their own account. The platform cannot do
it for them and must not try — the alternative is lending them the owner's
credential. claudeLoginState only reports whether the credential has appeared,
and reads it as the member, so a true answer means their process can reach it.
NOT VERIFIED: any of it at runtime. tsgo passes; nothing has been provisioned
and the spawn hook has never been called. If it turns out setpriv breaks how
the SDK reaches the process, this approach is wrong and the fallback is the
earlier plan.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A new Linux account opens a shell with nothing: useradd copies /etc/skel, which on Ubuntu
is a bash rc, and the account's shell is zsh — so it got no prompt, no history, no
completion, no colour. "Their own account" should not mean a worse terminal than the
owner's.
src/servers/shell-skel/zshrc is the template, and scripts/starship.toml is reused rather
than copied: setup.sh already deploys it for the owner, so one file serves both audiences
and they cannot drift. Seeded by provisionOsAccount, which means the retry button applies
it to accounts that already exist — no delete-and-recreate.
The template depends on nothing but zsh. Starship, eza, nvim, bun, deno and cargo are each
used only if present, and every path is $HOME-relative — the owner's own .zshrc has three
absolute /home/pastilhas paths in it, which is exactly what a template must not inherit.
Without starship it falls back to a zsh prompt showing the same information, because a
shell that opens with a broken prompt reads as a broken machine.
Never overwrites: written only when the file is ABSENT. ~/.zshrc.local is sourced last and
never written, so there is somewhere to put your own config that no future template can
reach.
Three fixes found by running it:
- install -D creates missing parents but applies -o/-g only to the FILE, so ~/.config came
out root:root — readable but not writable by its owner, which would have surfaced weeks
later as one tool mysteriously failing. The parent is now created explicitly.
- useradd took its shell from process.env.SHELL, which under PM2 is whatever PM2 was
launched from. A member's shell depended on how the server happened to be started. Now
chosen from what is installed: zsh, else bash.
- the pty sidecar spawned ITS $SHELL for a member, not theirs. It now execs their passwd
shell via sh -c, so the login shell in /etc/passwd is the one they get.
starship moves out of the light-profile skip. The light profile exists to serve a file
browser, a terminal and chat — the terminal is one of its three reasons to be, and it is
what every member gets. Leaving starship out meant the fallback prompt on exactly the
installs most likely to have members. oh-my-zsh, eza and lazygit stay full-only.
Verified in a real member shell: zsh from passwd, HISTFILE in their own home, eza-backed
ll, starship active, EDITOR=nvim, and an edit to .zshrc surviving a reprovision.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
TERMINAL is confined now, and the shell is genuinely theirs. The pty sidecar spawns it
through sudo setpriv as their own account, in their own home, with the platform's
environment cleared. Verified end to end against the sidecar's own socket:
id -u 1001, not 1000
file the shell wrote owned by ptyprobe
ps -o user=,args= ptyprobe /bin/zsh -i
env | grep -c POSTGRES 0
osUser and home are resolved in upgradeWs from the authenticated account, and whatever
the browser sent under those names is DELETED first. The bridge forwards the query string
to the sidecar untouched and the sidecar starts a shell from what it finds there, so
trusting the client for either would let a member ask for the owner's uid in a query
parameter.
node-pty does support uid/gid, unlike Bun.spawn, and they are deliberately unused: they
set the ids without applying the account's groups or resetting the environment, so the
shell would keep the owner's groups and everything Bun loaded from .env.
Also closes the pty identity blindness in TODO.md. Sessions record whose they are, list
and kill scope to the caller, and re-attaching to a session belonging to another account
is refused — otherwise a member resumes someone else's shell by guessing an id that
travels in a query string. Measured: member killing the owner's session -> ok:false,
owner killing it -> ok:true.
CHAT is confined so the owner can grant it and the route resolves, and both execution
doors refuse a non-owner: the router wholesale, and the socket in server.tsx. The agent
has not moved — the SDK spawns claude itself with nowhere to put a uid, and every
transcript path resolves through the owner's home, so a member would read the owner's
session list and run an agent as the owner. Reads are refused too, because
listClaudePwds returns the names of the owner's projects.
A deliberate, temporary gap at the owner's request: permission and route now, function
when a turn can be spawned under runAs with the member's own HOME. Both guards say so,
and the registry test names them so a future edit cannot move one without the other.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The serve is now the only way an opencode turn runs. runner.ts and its tests are gone, and
so is the OPENCODE_TURNS switch — there is no fallback engine any more, and the recovery for
a bad day is git rather than a config flag. Deliberate, and cheap right now precisely because
nothing depends on opencode yet.
What goes with it: mapRunLine and its NDJSON fixtures, the temp-file spill for --file image
attachments, the supersede-and-kill dance, the process watchdogs, the pidfile-adjacent child
tracking, and stopAllOpenCodeTurns. All of it existed to work around stdin being /dev/null.
Verified after deletion, with no env var set at all: tool call, tool result, 5 streaming
deltas, text and cost, through the real chat socket.
Also corrected the comments the deletion falsified rather than leaving them to mislead — the
module header, the wire contract description of opencode:run-streaming, and serve-runner own
header, which still announced itself as off by default.
One difference worth stating: shutdown no longer kills anything. Turns run inside the serve,
which is a separate process that survives us, so officer stops routing them and says so in
the transcript. When a subprocess ran the turn, failing to kill it orphaned it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
runOpenCodeTurnOnServe ignored params.images entirely, which is defect B4 rebuilt on the new
path: the image renders in your own bubble and the model never receives it, with nothing
reporting a loss. Fixed before the path is switched on for anyone rather than after.
data: URIs, not file://, and that is measured — the wrong choice is accepted with a 200 and
then dies inside the turn with "Anthropic Messages media must contain valid base64". The
data URI round-trips and the model describes the image.
Strictly better than the subprocess path here: no temp file to spill and nothing to clean up,
because the bytes travel in the request.
Both prompt paths carry them — an ordinary send and a mid-turn injection. Verified end to end
through the chat socket with the serve engine on: a red png came back "**Red**", with deltas
streaming.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Phase C server half, and a real flaw in phase B.
On the subprocess path a second message could only supersede — kill the process, start again,
lose the turn — because opencode run has no input channel. The serve takes another prompt
into the running turn, so a message arriving mid-turn is handed over with delivery steer and
the existing turn is left exactly as it is.
Keeping the same turn object is the load-bearing part. Phase B retired it and registered a
replacement, which stops officer routing events the serve is still producing while the serve
carries on regardless: output goes nowhere and the turn looks hung.
Verified end to end through the chat socket — sent a count to 50, injected a change of plan
eight seconds in, and BANANA INJECTED came back inside the same turn with deltas streaming
throughout.
No client change was needed. Officer composer already sends while generating; the difference
is only what the sidecar does with it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
OPENCODE_TURNS=serve picks the new engine; unset keeps the subprocess, which is the default
and stays the default until this has been lived with. A bad evening should cost one restart,
not a revert. Claude is a different sidecar and is untouched.
Verified end to end through the real chat socket:
session:init -> tool:start(bash) -> tool:result -> assistant:delta x3 -> assistant:text
-> result, cost in=304 out=73
Those deltas are the first token streaming an opencode turn has ever produced in officer.
Stop is now an INTERRUPT: the turn ends and the session survives — verified by sending a
second prompt to the same session afterwards and getting an answer, which killing a
subprocess could never do.
Reads the LIVE global stream rather than the durable per-session one, because it is a strict
superset — same tool.called, tool.success, step.ended, text.ended, plus the deltas that are
the whole point. Global means one socket carries every session, so everything filters on
sessionID; one subscription is shared for the process rather than one per turn.
A turn ends on step.ended with finish != tool-calls. tool-calls is a step boundary MID-turn,
and treating it as terminal would cut every tool-using conversation in half.
delivery is stated explicitly as queue because it DEFAULTS to steer, which injects into a
running turn — wrong for an ordinary send, where two quick messages would merge into one.
Wiring steer to the button that means it is phase C.
What phase B does not do: read the durable stream. The sidecar still commits every event to
chat_session_events as it arrives, so durability is unchanged, but recovering a turn this
process never saw needs the ?after= cursor and is its own change.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The mapping half of the serve migration, written and pinned before anything depends on it,
so the switch-over is not also the moment the parsing turns out to be wrong. Nothing routes
through this — turns are still opencode run subprocesses, and the claude path is untouched.
The finding that matters: the serve publishes each turn TWICE, and reading the wrong one
makes it look like it cannot stream at all.
/api/session/{id}/event?after= durable, per session, replayable, durable.seq on every
event, whole values only, NO deltas
/api/event live, GLOBAL, ephemeral, carries text.delta and
tool.input.delta, no cursor
Same turn: 13 events durable, 21 live, the difference being 3 text.delta and 5
tool.input.delta. I probed the per-session one first and nearly recorded "no streaming" as
a fact — it would have removed the main reason to migrate. The split maps exactly onto what
officer already does for claude: durable to chat_session_events, live to UI deltas. The cost
is that the live stream is global, so a consumer must filter on sessionID.
tool:start is emitted on tool.called, not tool.input.started, because only tool.called has
the resolved input object — the input arrives as JSON fragments ({"comman) and a tool row
rendered with half-parsed arguments is worse than one that appears a moment later.
step.ended with finish tool-calls is a step boundary MID-turn, not the end of the turn, so
nothing terminal is emitted for it. Treating it as the end would cut every tool-using
conversation in half.
Fixtures are verbatim captures from 1.18.16. Replaying both real streams through the mapper
reconstructs the turn identically from each, with the reassembled deltas exactly equal to
the committed text and identical cost, and zero unrecognised events.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Bun.spawn throws on a missing or non-executable binary rather than resolving to a failed
process, and that throw escaped runOpenCodeTurn entirely — past the bookkeeping, out of the
sidecar command handler, with no opencode:event ever emitted. The browser sat on a spinner
nothing could end, because the code that ends turns had not been reached.
A wrong OPENCODE_BIN is the ordinary way to get there, so the message names the path it
tried: that is the difference between a fix and a debugging session.
Also records that messageCount is not a gap. SessionList renders an OpenCode badge in place
of the count for those rows, so the hardcoded 0 never reaches a screen, and computing a real
one would cost an HTTP call per listed session — the session record has no count field — to
populate something nothing shows.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
opencode keeps credentials in two unrelated places. The CLI, opencode run and the legacy
/session surface read auth.json. The newer /api surface — the one with steer, queue,
interrupt and a resumable per-session stream — reads its own integration store and knows
nothing about that file.
With none connected it does not fail. It falls back to what needs no credential, the free
tier, and a request for a paid model is never executed: prompt accepted, admitted, prompted,
then no step, no error, no message, forever. That silence cost most of an afternoon and would
cost it again on every new machine — alpha included.
So the sidecar does it, rather than depending on someone having run a curl. Best-effort and
never blocking: turns go through opencode run, which reads auth.json and does not care.
Retried, because /api/health answers before the integration store is ready — the first
version of this shipped without a retry and failed on its very first real boot with a 500,
while the identical request succeeded seconds later. Only 5xx retries; a 4xx means the
request is wrong and repeating it just prints the same complaint six times.
Verified by deleting the credential, restarting, and running sonnet on the new pipeline with
no manual step.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
B4 properly. The composer gate was the honest stopgap; this is the fix. opencode run takes
attachments with --file, so images work on the subprocess path we already use — the parity
doc had them down as phase 4, behind the serve migration, and they were not.
The bug was one omission: handleOpenCodeChat`s msg type had no images field, so the browser
sent them, the bubble rendered them, and they stopped at that signature. Nothing reported a
loss anywhere.
Attachments are paths, not inline data, so the sidecar spills each image to a temp file for
the length of the turn and removes it in settle — the same place every other per-turn
resource is released, so a killed or superseded turn cleans up too.
The load-bearing detail is `--` before the prompt: --file is an array option, so without the
separator the prompt is eaten as another filename and the turn dies with "File not found:"
followed by the entire message. Confirmed against the binary, and pinned by a test that
records argv from a stub.
list-models now reports each model own capability instead of a hardcoded false — opencode
publishes capabilities.input.image per model and nothing had ever read it. Defaults to false,
so a model that does not declare it keeps the affordance hidden.
Verified end to end: a red png sent over the chat socket to opencode/claude-sonnet-4-6 came
back "Red".
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The previous commit migrated only when SQLite was completely empty, on the assumption that a non-empty
file is an authoritative one. A real account disproved that within minutes of it landing.
The older Gmail backfill had already written SOME keys into SQLite — last_sync_at and the uidvalidity
set — and never the imap_lastuid ones. So the file was non-empty and half-migrated at the same time,
all-or-nothing skipped the migration, and nine imap_lastuid keys stayed only in Postgres. A missing
lastuid makes the next sync refetch that folder from UID 1: on the mailbox this was found on, 18,755
messages and 6.9 GB.
Now merged per key with the file always winning a conflict. That keeps the property all-or-nothing was
protecting — a restored older emails.db still overrides a newer Postgres row for every key it has, so it
cannot be advanced past mail it does not contain — and adds the keys the file never had, which are
exactly the ones whose absence is expensive.
Verified against the live account: all 22 Postgres keys present afterwards, imap_lastuid:INBOX restored
to 208407, and last_sync_at left at the file's older value, so it re-checks a fortnight rather than
skipping it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The messages were in emails.db and the position — last_sync_at, and per-folder uidvalidity/lastuid —
was a jsonb column on email_accounts in Postgres. Two stores for one fact, with an edge that only shows
up when you try to move a mailbox to another machine.
The expensive part of an email account is the first sync: hours of IMAP for a large mailbox, which is
exactly why "copy emails.db to the new server" is the obvious way to bring one across. With the position
in Postgres that silently does not work — the new server's column is empty, !last_sync_at says first
sync, and the whole mailbox downloads again on top of the one just restored.
The other direction is quieter and worse. Restore an OLDER emails.db while Postgres holds a NEWER
position and the sidecar skips every message between the two, permanently, because nothing looks below
lastuid again. Re-syncing is slow; skipping mail is data loss nobody notices.
Not a new idea — the Gmail path already read SQLite and fell back to Postgres, backfilling so the
fallback was taken once. Only the IMAP path had not followed. This extracts that pattern so both use one
copy, and unifies the isFirstSync fork in accounts.ts, which is how the two drifted apart to begin with.
The file wins over Postgres, always, and only migrates when it holds nothing at all. Topping up a
partial position from Postgres would reintroduce precisely the divergence this removes.
email_accounts.sync_meta is kept and marked legacy rather than dropped: it is the one-time backfill
source for every account created before this, and dropping it would strand any that has not synced
since. Nothing writes to it now.
11 tests on the migration, aimed at both expensive failures — migrating when we should not, and failing
to migrate an account that predates the change.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The serve has a second, newer API surface nobody here had looked at, and it publishes
exactly what the parity doc calls impossible under stdin ignore: delivery steer and queue
on POST /prompt, an interrupt that does not tear down, and a per-session event stream with
an after cursor — the durable-replay machinery officer hand-built for claude, as a
primitive. That would have made migrating obvious.
It does not execute. A prompt is accepted with an admittedSeq, stored, emits
prompt.admitted and prompted, and then never steps. Ruled out separately: the model, the
permissions (build is *:allow, no pending requests), the per-request location (the surface
is location-scoped via header or a deepObject query, supplied everywhere, no change), and a
config gate. The legacy POST /session/id/message?directory= generates fine in 17s, so the
serve itself works — only the new pipeline is inert. session.next.* is the tell.
And not a version problem, which is the part everything here had backwards: this Mac runs
1.18.11 and alpha runs 1.17.9, measured. The dead pipeline was tested on the NEWER binary.
The original "this server runs 1.17.9" meant alpha and was copied to a machine where it was
false; corrected in runner.ts and the test.
So building against it now would produce code that looks finished and does nothing, which
is the failure mode this project keeps rediscovering. One request reopens the question
after any upgrade, and the doc names it.
Also de-flakes the lifecycle tests: they spawn real processes, and a fixed sleep(750) went
red once on a machine busy running these probes. Presence assertions poll now.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
They were permanently unnamed, and the two halves needed to name them already existed on
officer: the sidecar reports its own sessionKey because that is all it has, while the ses_ id
arrives separately over opencode:session and is recorded in opencode/state.ts. Nothing joined
them. /chat/live joins them now, so no protocol or sidecar change — widening
LiveOpenCodeSession would have meant sending the sidecar a fact it told officer in the first
place.
One list call names every row rather than one transcript load each, and it is skipped when
nothing is running or no id has been reported, so an idle Live panel never touches the serve.
Verified against a real turn, which also showed the design working as intended: the first
poll has no id yet and shows nothing, the next shows title and cwd. That window is real and
short, and showing nothing beats showing a key the user has never seen.
Worth knowing: opencode titles its own sessions "New session - <ISO timestamp>", so the row
is located but not meaningfully named. That is genuinely its title, not a bug here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
B8, both halves. They share index.ts, so they share a commit.
In-flight turns: `opencode run` is spawned, not supervised, so pm2 restart officer-opencode
left every turn ALIVE — reparented, still spending tokens, still writing files as the agent,
with the only reader of its stdout gone. The transcript stopped mid-tool-call, which reads
as the agent hanging.
stopAllOpenCodeTurns kills them and settles each synchronously, because the caller is about
to process.exit and nothing waiting on proc.exited would ever run. Settling writes a reason,
so a reload after a restart explains itself instead of trailing off. Turns are stopped BEFORE
the connection is destroyed — that write travels over it — and the flush is bounded, since
losing the explanation is bad but hanging the restart is worse.
Stale serves: the sweep read /proc, so it was a no-op on macOS and orphaned serves piled up,
one per unclean exit, each holding a port. Added a pidfile sweep alongside it. A pid we wrote
ourselves needs no cwd guard to prove it is ours, which is the part ps cannot answer portably
(macOS would need lsof), and a serve started by hand is never in the file.
The guard checks command AND subcommand: matching the word serve anywhere in the line would
sweep a running turn whose prompt merely mentioned it. Fixtures are real ps output from both
machines, not invented. Split into serve-sweep.ts because index.ts spawns a serve at module
scope, so a test importing it would start one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Two comments describing behaviour the code does not have.
LiveOpenCodeSession had been inserted between LiveClaudeSession and its docblock, so a
comment about isGenerating, pendingTasks and the idle GC read as documentation for the
OpenCode type — where it is contradicted by the correct comment directly beneath it.
Moved below, and it now states that it carries no ses_ id.
That absence is the point: /chat/live claimed title and cwd come from the session store
"so a turn whose id has not been reported yet shows unnamed". Nothing is looked up, and
there is no id here to look one up with. They are null permanently, not until-known.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A replaced turn is killed but dies asynchronously, so its proc.exited fired long after
the replacement was registered under the same sessionKey — and then ran the whole
completion path against it: emitted "OpenCode exited with code 143", which the sidecar
commits to chat_session_events so a false failure became permanent history, then deleted
the replacement from `running`. That blinded the new Live panel, made the stop button a
no-op and orphaned a process nothing could reach.
Mark the handle before killing it, retire it silently, and identity-check the delete —
a superseded turn does not own that key any more.
Two leaks in the same family, found while fixing it. An early return would not have been
enough: both watchdogs call finish, so the armed 10-minute hardTimer would have fired an
error at whichever turn held the key by then. And handleLine had no `done` guard, so
stdout still draining from the killed process was emitted under the replacement key.
Reproduced before fixing. The lifecycle tests need no real opencode — RunnerConfig.bin
takes a shell script that sleeps. The control test pins that an ordinary non-zero exit
still reports an error, so the guard cannot overreach.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The Live panel asked `claude:list` and nothing else, so a running OpenCode turn was invisible — the panel
claimed to show what the agent is doing and silently omitted half of it.
Adds `opencode:list` / `opencode:sessions` and merges both harnesses in `/chat/live`, asked in parallel,
each failing toward empty so one sidecar being down contributes nothing rather than breaking the panel.
The OpenCode row is deliberately thinner than the Claude one rather than faked into parity:
isGenerating always true — a subprocess exists only while it generates, so there is no "merely open"
pendingTasks always 0 — `opencode run` has no background-task concept; reporting a number would
suggest a capability that does not exist
title / cwd null — the session store is keyed on the `ses_…` id the runner reports, not on
our sessionKey, so an unreported turn shows unnamed rather than guessed
This is the incremental option from docs/opencode-serve-path.md — enumeration without moving turns onto
the serve, so it buys the Live panel with no warm sessions, no SSE loop and no lifetime questions.
NOT verified end to end: no OpenCode turn was running to enumerate, so the verb is wired and typechecked
but has never returned a non-empty list. See COMMS/BLOCKERS.md.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Path names: the serve's cwd is DATA_PATH/opencode_server, not opencode-sidecar. Two comments said
otherwise and would send the next reader to a directory that does not exist.
Version pin: the comment claimed "verified live against 1.17.9" as though that were a property of the
code. It is a property of whichever binary is installed, and this project already runs two — 1.17.9 here,
1.18.11 on the other machine. Says so now, and points at the test as the thing that actually enforces it.
Tests, the first on the OpenCode path. `runner.ts`'s NDJSON → ChatEvent mapping was described as pure and
untested; it was untested but not pure — it lived inside `handleLine` as a closure over `emit`, the
accumulated cost and a reported-session flag, so it could not be called without spawning a binary.
Extracted as `mapRunLine`, genuinely pure: line in, {sessionId, events, costDelta} out. The two concerns
that span lines stay with the caller, because they are not properties of a line — emitting the session id
exactly once, and accumulating cost across steps. Behaviour is unchanged.
11 tests over what the mapping forwards, what it drops and what it must not turn into NaN. The last one
matters: a missing `cost` on a step_finish would otherwise propagate NaN into the turn total.
Phase 1 is complete: dead code deleted (previous commit), comments corrected, tests added.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The sidecar seeded an AGENTS.md into the serve's project root telling the agent to read its working
directory from the "Working directory for this session" line in its system prompt. The run path sends no
system prompt, so there was no such line and the instruction had been inert since turns moved off the
serve to `opencode run --dir`.
What it was standing in for, `--dir` does properly — tested rather than assumed
(docs/opencode-phase0-review.md): `--dir` anchors the agent's own file operations, not just the process
cwd, and the anchor survives a multi-step turn with a write in the middle. Nothing replaces it.
The generated file is removed from disk too, not only from the code that wrote it; leaving it would have
kept feeding standing instructions to every session while looking, in the source, as though it were gone.
Also corrects the claim that all OpenCode sessions live in one server's project. True when turns
inherited the serve's directory, false now: one serve lists 7 sessions across several directories, which
is why the cwd filter has to read `directory` rather than assume a single one.
docs/opencode-phase0-review.md, item 2 — Phase 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
An upload over a few MB came back as a 400 with an empty body, no message anywhere, and nothing logged
by Immich, the sidecar or officer. It was a race, not a size limit. Immich judges an asset from its
first few KB and rejects immediately, then closes; both our hops were still writing the body; Node
treats the leftover bytes as a protocol violation and replaces the application's answer with a bodyless
`400 Bad Request` + `Connection: close`. The real message never reached the wire.
Measured before the change: streamed lost the message 1/4 at 8 MB and 4/4 at 32 MB — probability rising
with size, which is why small photos usually worked and a phone's video never did.
Both hops needed it. Fixing only the sidecar took 32 MB from 4/4 failing to 2/4, because the platform
proxy was losing it one hop up.
Bounded at 512 MB, above which the body streams exactly as before. That ceiling is not a refusal and is
deliberately not a 413: a file Immich ACCEPTS is read to the end and never races, so a 4 GB video is
unaffected. All that is given up above the cap is the error message on a file that was going to be
rejected anyway. A first attempt refused over-cap uploads outright and would have broken the working
4 GB case to improve diagnosis of the doomed one.
`bufferRequestBody` is opt-in and off by default: the vault and wallet proxies must keep streaming so a
passphrase or macaroon never lands in the platform's heap.
Also adds the proxy error logging that made this findable at all — status and two byte counts from
headers, never the bodies. `responseBytes: "unknown"` is what exposed the stripped response.
Verified live at 32/256 MB (buffered) and 640 MB (streamed, passes through).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every live row read 'Starting…' and linked nowhere, because the title lookup searched for a transcript
named after the session KEY. It isn't. The key is officer's handle for a conversation; the transcript is
named after Claude's own session id, and the mapping between them exists only inside the agent
(setClaudeSession/getClaudeSession). I assumed the two were the same and never checked — confirmed wrong
by looking for the ids from the officer log under ~/.claude/projects and finding nothing.
claude:list now reports claudeSessionId beside the key, the route resolves titles by that, and rows link
to it. Null means the first turn has not reported one yet, which is a genuinely unwritten conversation
and stays unlinked.
Needs the agent sidecar restarted to take effect — the new field comes from there.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
resolveOwner already retried forever — but only when the query SUCCEEDED and returned nothing, which is
a fresh install waiting on bootstrap. A query that THREW escaped the function, rejected the top-level
await and exited the process, into exactly the PM2 restart loop its own comment says it exists to avoid.
So any Postgres restart (57P03 'the database system is starting up') or moment of unavailability killed
every live agent session on the machine and spun the sidecar until the database answered.
That is what took a session down on 2026-08-10, and why this process showed 468 restarts against 0 for
every peer that starts without needing the database.
The loop now catches as well as checks. Still retries forever, matching the case beside it: a database
coming back is a matter of time, and an agent that gave up would need a human to notice.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Officer's session records are in memory and die with `pm2 restart officer`, while the agent is a PM2
peer and keeps generating. `adoptOrphanedSession` rebuilds a binding — but only when a browser
reconnects to a session *by id*, which you can only do if you already knew the id. So a session that
survived a restart was invisible, and nothing could answer "what is running right now".
`claude:list` returns each live session with `isGenerating` and `pendingTasks` — the same two fields the
agent's own `armIdle` consults before collecting a session, so a caller can tell "busy" from "merely
open" the way it does. Surfaced as `GET /chat/live`, which sits beside `/chat/sessions`: those are
transcripts on disk, these are the ones with a process behind them.
`getActiveSessionKeys` is replaced rather than joined. It returned bare keys, could not distinguish a
session mid-turn from one merely open, and had never been called by anything.
`listLiveClaudeSessions` fails toward EMPTY, where `isClaudeGenerating` beside it fails toward alive.
The asymmetry is deliberate: not knowing there means leaving a spinner up, and not knowing here would
mean inventing sessions.
Step 2 of docs/chat-session-lifetime.md.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Platform chat ran at 200K while the same `claude` in a terminal got Opus 5's
full 1M. Nothing to do with the model, the account or compaction tuning — the
CLI gates 1M on `provider === 'firstParty' && Fp()`, and `Fp()` is satisfied
only when ANTHROPIC_BASE_URL is unset or its host is api.anthropic.com. We
point it at 127.0.0.1:5051 so the agent's traffic goes through the OAuth
proxy, which fails that host check and silently drops the window to 200K.
_CLAUDE_CODE_ASSUME_FIRST_PARTY_BASE_URL is the CLI's own escape hatch for a
first-party passthrough, and the assertion holds: the proxy forwards verbatim
to api.anthropic.com and already preserves anthropic-beta.
Read from the CLI binary (2.1.223), not inferred: claude-opus-5 carries
context:{window:1e6, native_1m:true}, so the model half of the gate always
passed. Only the hostname was wrong.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
both forwarders buffered the whole request body into an ArrayBuffer before
re-sending it to Immich. with maxRequestBodySize at 4GB that put a phone's
video upload in the sidecar's heap for a hop that never reads the bytes.
callUpstream now sets duplex: 'half' so a stream is a legal body, matching
what createSidecarProxy already does on the platform side. the four JSON
callers are unaffected.
the platform proxy forwards no content-length, so the body already reached
us chunked; this extends that one hop to Immich. verified against the live
instance (3.1.0): bulk-upload-check round-trips a streamed body and returns
the right verdict.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Locked assets are gated on elevation, and elevation lives on a session — an API key
has no `auth.session`, so no key permission or allow-list entry can reach them. The
gate sits inside the generic owner-access check, so a locked asset's thumbnail and
original are covered too, not just its listings.
So `/_locked` mints a session at unlock, holds it in memory for the elevation window,
and closes it on lock, on idle, or when the active immich account changes. Nothing new
is written to photos_config: the pin and the password are never at rest, and a full
compromise of officer's database still does not open the folder. The cost is that
unlocking asks for the immich password as well as the pin.
`auth/*` stays refused wholesale in routes.ts. The four auth routes this needs are
reached through named endpoints that each do one thing, and the elevated forward
carries three resources rather than the main allow-list.
Not yet exercised at runtime — the sidecar has not run this code.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Extracting resolveNotifyUser into its own module immediately caught a hole
in the fix from the previous commit: a header that was present but
unparseable fell through to the body, so a browser could send junk in the
header, name any user in the body and win.
PRESENCE of X-Officer-User is the signal, not its validity — a malformed
header means a proxied request went wrong, and falling through hands the
decision back to the caller we just declined to trust.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A turn could stop producing events and stay `isGenerating` indefinitely. Nothing
covered it: the idle timer answers the opposite question — how long a session with
NO turn in flight may sit before collection — and every client showed a spinner
with no timeout of its own, so a wedged turn presented as a chat that was still
thinking.
On 2026-08-08 one ran for seventeen minutes inside an auto-compaction, reached
over the socket to an iPad, and was indistinguishable there from a dead app. The
compaction is silent by design (the PreCompact hook is the only announcement, and
the code's own comment allows 2.5 minutes), so there was nothing to distinguish it
from.
A stall watchdog now rides every emitted event: any sign of life pushes the
deadline back, and expiry ends the turn the way a real failure would — isGenerating
off, idle re-armed, and an `error` the client can render. The agent process is
deliberately left alive, since it may still be working and the next turn resumes
it; what this guarantees is that the client is TOLD, which is the part that was
missing.
The budgets are generous rather than tight — ten minutes of silence normally,
twenty while compacting, re-armed from the PreCompact hook because that hook fires
as the long silence begins and the deadline the turn is holding was sized for
ordinary work. Killing a turn that was about to succeed is worse than the hang this
prevents.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
On a Mac, Claude Code stores its credentials in the login Keychain and never
writes ~/.claude/.credentials.json — the only file this proxy knew how to read.
The workaround was to copy the Keychain blob into that file by hand, which is a
snapshot: a refresh ROTATES the refresh token and revokes the previous one, so
the two stores were not redundant copies but competitors, and whichever
refreshed second got `401 OAuth access token has been revoked`.
That is not hypothetical. On 2026-08-08 it took out every chat turn from the
iPad for six hours while the terminal CLI beside it worked fine — the harness
spawned, retried for three minutes and wrote the 401 into the transcript, which
from the app looks like an agent that simply never answers.
So on darwin the Keychain is the authority and the file is a mirror, holding the
same token rather than a different rotation of it. Everywhere else — every Linux
server — the file is still the authority and nothing changes. Detection is
process.platform, and a machine with no `security` binary or no such item falls
through to the file rather than failing.
Three recoveries, cheapest first:
- a watchdog checks every 30 minutes and refreshes when under an hour remains.
It checks rather than refreshing on a blind schedule because each refresh
rotates the token, so a needless one is another chance for the stores to
disagree.
- an upstream 401 now RE-READS before refreshing. When a token has genuinely
been revoked the machine usually already holds a good one, because Claude Code
refreshed it into the Keychain minutes ago; spending our own refresh token
there is what caused the divergence in the first place.
- only if nobody else has moved do we refresh ourselves.
The Keychain write goes through argv, which is the only non-interactive form
`security` offers, and matches on the service AND account pair — the account is
read off the existing item rather than assumed, or the update would silently
create a second entry instead of replacing the one Claude Code reads.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
CLAUDE.md asserted "single-user is a hard invariant, not a stage" while
users held six rows and role_capabilities held grants. Every doc that
repeated it is corrected here, in prose and in the code comments that
carried the same claim.
The accurate statement is narrower: one owner who bypasses every check,
other accounts holding only what their role is granted, and a set of
capabilities — terminal, chat, files, tasks, items, desktop, browser — that
are structurally ungrantable because they execute as the owner's OS user.
TODO.md gains a Multi-user section for what the read turned up: no way to
create a second account, dashboards.id colliding across users, authorize.ts
untested, pty/vault/opencode taking no identity, Radicale still owner_only.
claude-sidecar-isolation.md's open question is answered rather than left
open — the per-email spawn model is dead weight, because chat is an
execution capability and no second account can ever reach it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
reset-password accepted any valid signed jwt as a reset token, including a
30-day session token — its sibling verify-token.ts already gated on
purpose === 'reset-password' and this handler did not. forgot-password mints
that claim, so the gate costs the legitimate flow nothing.
notify's DELETE /_officer/devices/:token deleted by token with no user
predicate: a token is the address of a device, not a secret, so any account
holding the notify capability could deregister another's device.
deletePushDevice now takes an optional userId — the route passes it, the
APNs/FCM dead-token paths deliberately do not.
POST /_officer/notify let a request body's userId override the
proxy-injected X-Officer-User. The header now wins where present, which is
what separates a signed-in browser from a loopback producer that has no
session to speak from.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
adding an email account failed with a bare "failed to add account" toast. three
defects stacked, each hiding the next.
the email sidecar's http.ts reconstructs what the platform's middleware used to
provide, but only did two of three — bodyParser was never remounted, so every
write route read ctx.get('body') as undefined and POST /accounts threw on
body.provider before ever reaching the credentials.
its onError then read `.status` off the thrown custom-error, which carries
`statusCode`. every deliberate 4xx fell through to the 500 branch and had its
message replaced with "internal error", so a rejected IMAP login and a genuine
crash looked identical. it also answered JSON where the rest of the api answers
errors as plain text. now mirrors hono.ts's handler rather than inventing a
second shape.
useClient threw a plain object, so the ~33 sites narrowing with
`err instanceof Error ? err.message : <fallback>` always took the fallback and
discarded the server's message. now throws an ApiError subclass keeping both
status and message, so those sites start surfacing real errors.
only email reads ctx.get('body'); every other sidecar is a pure proxy.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Refreshing mid-turn appeared to kill the agent's output. It never did: the
session survives a dropped socket, the agent keeps generating into it and keeps
committing durable events, and `close` only detaches the socket and arms an
hour-long idle timer. What broke was purely delivery — and the reconnect path
that would have fixed it could not fire, because the browser came back having
forgotten officer's session key. It lived in page state. The only id left was
Claude's transcript uuid in the URL, and nothing accepted that.
So accept it. `attach` carries the uuid, and the agent's on-disk session map —
the single record relating the two — turns it back into the key everything else
is written in terms of. The uuid now also goes out at `system.init` rather than
only at `result`, which is what makes the first turn recoverable at all: until
now a chat had no address until it had finished, and a long first turn is
exactly the one worth reconnecting to.
`sync:live` deliberately carries no messages. The harness writes its transcript
as it goes, so the HTTP load on landing already supplies the past; sending the
server's record of the same messages on top of it would duplicate them, and
there is no shared id to reconcile the two by. Attach hands over the rest of the
turn, the half-written paragraph the transcript cannot hold, and the session's
cursor head — that last one so a *later* drop replays from the head instead of
re-delivering the whole conversation from zero.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Compaction was the one thing the harness does that emitted nothing at all while it
ran, and it can run for minutes — silence that reads as a hung turn, which costs a
server restart to discover it wasn't. The sidecar now reports both ends: the start
from the PreCompact hook, the finish from the compact_boundary message with the
token count, both durable so a reload or a reconnect still sees them.
Tool rows open themselves while they run and hold for five seconds after their
result, so the inputs are on screen at the moment the call is made rather than
after the fact. The clock lives outside React, keyed by tool call id: rows are
virtualised, so unmounting is not the call ending, and a fast call can render its
start and its result together — a row that only opens when it catches the pending
state never opens for exactly the quickest calls. A click outranks the clock for
as long as the row lives. A failure behaves identically and differs only in colour,
so it stays findable by scanning and nameable in conversation.
Shell logs move off the green-on-black pre onto the shared code surface, which is
the one block that had no copy button and the one you most often want to hand to
someone else.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
every multipart upload through /api/<sidecar>/* arrived corrupted. bodyParser ran on
proxy routes and called parseBody for multipart, so hono cached a FormData on the
request; when the proxy then asked for the bytes hono re-serialised them from that
cache with a NEW boundary, while the proxy still forwarded the ORIGINAL content-type
header. header and body disagreed and the far side rejected it with
"Multipart: Unexpected end of form".
bodyParser now skips prefixes owned by createSidecarProxy, which register themselves
so a new sidecar cannot forget. the proxy also forwards the body as a stream instead
of buffering it, which drops the second in-memory copy of every upload.
note the bug report proposed skipping multipart in bodyParser outright; that would
have broken /upload, /file-browser upload and /bug-report, which do read a multipart
body from ctx.get('body').
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The composer already uploaded an image, split its data URL and put the bytes on the wire as
`images`. Nothing on the server read them. The `chat` ClientMessage had no such field, and the
prompt reached the sidecar as a bare string, so all the model ever saw was the client-generated
`[Attached image: …]` placeholder — a label describing a picture it was never shown.
The transport was never the obstacle: `query()` consumes an async iterable of user messages whose
`content` is an Anthropic `MessageParam`, and only `pushTurn` hardcoding a string kept it to text.
So `images` is threaded through the four hops that dropped it and turned into native image content
blocks at the end, renaming `mediaType` to the API's `media_type` at that last step.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
restarting officer-agent takes every persistent session with it and nothing
downstream notices: the browser's socket is healthy, officer's subscription is
a bus filter, and there is simply never another event. the spinner ran forever
and a refresh didn't help, because the transcript has no ending to read.
keyed off the agent *registering*, not disconnecting — a disconnect fires on
every `pm2 restart officer`, when the turn is fine. a registration socket dies
with its process, so an agent appearing on it is a new one. covers the sitting
tab; the reconnect path covers the rest, with the client now sending its belief
that a turn is in flight and officer checking it against the agent over a new
claude:is-generating. the check fails toward alive.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
pastilhasandClaude Opus 5