✔ every screen and overlay fits 120x36, 80x24 and the narrow layout, row for row (2438.749054ms)
✔ width counts terminal cells: wide CJK and emoji 2, U+FE0F after a one-cell base 1, marks 0 (117.357779ms)
✔ broker text is shown, never interpreted: no raw control character or bidi override reaches the terminal (188.251063ms)
✔ NO_COLOR: no escape codes, except reverse video on the selected row (325.007892ms)
✔ 1-9 shows the decide flow's own lines and question; only y resolves (67.258113ms)
✔ the app shows the same decide text as mosaic decide (33.944924ms)
✔ any key but y or Y at the question answers No; nothing is sent (267.029533ms)
✔ below 40x10 the question is hidden, so every key answers it No, y included (Filbert R3) (105.271132ms)
✔ under 40 columns a tall terminal shows the too-small screen even where the decide box would fit; y sends nothing (Filbert N10) (352.417776ms)
✔ a hidden decide box names the way out, and the status line under it does not promise y (Filbert N12) (37.044235ms)
✔ a throw while answering the question goes to crash, sends nothing and leaves no job pending (25.067949ms)
✔ type-ahead never answers the question: a paste, keys typed before it is drawn, keys inside askMs (422.081433ms)
✔ quitting at the question answers No and sends nothing, armed or not (42.254804ms)
✔ a timer tick reads nothing under an overlay, never answers the question, and keeps a resolved status (41.273648ms)
✔ a long question is cut inside the question; the action, target, choice and prompt stay (Dewey B1) (58.726209ms)
✔ a decide box too big for the terminal says so, and every key answers it No, y included (Dewey B1, round 3) (367.227213ms)
✔ at every size the decide box shows each line y depends on, or says it is hidden (9051.788658ms)
✔ outcome unknown: the CLI's sentence, no retry, then the trail or a re-read (73.541435ms)
✔ refused and failed reads keep the last list, marked stale; a refused decide says refused (34.433395ms)
✔ help or a detail box cut to fit keeps its last line, Esc included (Dewey) (18.474143ms)
✔ keys: views, moving, trail of a selection, the command line, help and quit (99.315056ms)
✔ one option: the status line says Press 1, not 1-1 (21.21112ms)
✔ no business: the app says so and :business picks one (12.510739ms)
✔ mosaic with no argument: exit 3 in an agent run, exit 4 with the usage off a terminal, the app on one (2.273532ms)
✔ scripts/mosaic with no argument reaches the app entry; the seat usage it printed before has not drifted (204.296317ms)
✔ strings to keep exactly live once, in the decide flow and the transport; the app does not restate them (0.48696ms)
✔ decide with --note still asks; a No sends nothing (12.17771ms)
✔ the app's business comes from a live bus host only (1.226836ms)
✔ runApp leaves the terminal as it found it on q, end of input, SIGTERM, SIGINT, SIGHUP and an error, the first frame's included (8.32498ms)
✔ draw is pure: same state, same frame (0.839823ms)
✔ recordBootRefused keeps the last good load across refusals; recordBootOk clears the refusal (5.212109ms)
✔ an unknown key and a JSON syntax error each leave the bus down with the reason and path in the status (1345.524118ms)
✔ a missing file is recorded too, and a valid file clears the status on the next start (784.872012ms)
✔ CLI off, absent and other-only channel selections refuse every project/thread flow before any effect (23.667466ms)
✔ CLI enabled project/thread lists dispatch through selection without changing output (2.622129ms)
✔ both channels off recover through the existing human CLI, socket and organization writer (3023.497574ms)
✔ inbox lists only decisions routed to the human, with what approving authorizes and how to decide (46.171876ms)
✔ decide resolves by id prefix with --yes, and the trail shows it in broker order with the task to follow (31.963997ms)
✔ decide refuses without a terminal or --yes, on an unknown option and on a short reference (27.915421ms)
✔ decide prints a declining choice as declining (32.442793ms)
✔ an unknown outcome is reported once and never resent (24.840727ms)
✔ a decision closed before the answer arrives exits 2 and points at its trail (23.133018ms)
✔ a prefix that matches two open decisions exits 2 and resolves neither (24.920614ms)
✔ without --business a command uses the live host's business, and a stale host.json is not a host (33.399211ms)
✔ every human command refuses inside an agent run before it touches the bus (52.33396ms)
✔ usage errors exit 4; no business and no host is a usage error (29.246916ms)
✔ agents and tasks print through the broker (21.044362ms)
✔ notify.json: missing, loose, malformed and extra keys refuse; a binding or null passes (2.211602ms)
✔ bootConfig builds the broker's boot message for one business, with no trackers key when no project names one (72.912029ms)
✔ trackers come from the tracker.* variables of the one project that names a tracker project (114.183196ms)
✔ with two projects, the one that sets tracker.project supplies the tracker and the other is no conflict (63.77407ms)
✔ two projects that each name a tracker project refuse, since the boot shape holds one (58.387926ms)
✔ a business without tracker.baseUrl gets no trackers entry (97.460837ms)
✔ an unknown business and a broken system config refuse with exit 3 (102.823992ms)
✔ follow prints an unsolicited message after plain talk returned (58.105909ms)
✔ talk --follow continues after reply and prints a later unsolicited message (45.838888ms)
✔ inbox lists all unread human messages under decisions, does not deliver or read them (580.289438ms)
✔ follow refuses agent context and bad arguments (21.988471ms)
✔ follow Ctrl-C wakes the poll sleep and removes its handler; transport errors also clean up (22.091999ms)
✔ plain talk leaves the process SIGINT handlers unchanged (21.893313ms)
✔ empty views say so (0.995175ms)
✔ the trail keeps the broker's order and names a decision's task without its rows (1.320899ms)
✔ tasks print the tracker fields the snapshot carries (0.209666ms)
✔ the review loop: the coder pushes and asks, the reviewer fetches and approves, and the task's trail holds each step (806.536221ms)
✔ refusals: no action, non-fast-forward, an unknown commit or request, and a push for another task moves only that task's ref (902.47872ms)
✔ the size cap refuses a push before any byte (365.842333ms)
✔ a fetch over the cap refuses before any byte reaches the reviewer (496.605756ms)
✔ the socket refuses a request with no capability, an unknown one, and a bad request line (276.237107ms)
✔ a rigged clone runs nothing: hooks, fsmonitor, a filter, a gitlink and a symlinked .git/config leave no marker (452.720657ms)
✔ an identity with no role, or of another business, refuses not-a-role and moves no ref (295.709144ms)
✔ a review of a project the business no longer has refuses unknown-project (277.761318ms)
✔ past maxActive connections the socket refuses git-busy, and a hung-up connection frees its slot (194.627683ms)
✔ close() answers host-closing to a connection still sending its request line (183.630135ms)
✔ a bundle that trickles in refuses body-timeout at its deadline and frees its slot (615.854266ms)
✔ a reviewer gone before its bundle was sent is filed fetch-aborted, not as the host's fault (580.106724ms)
✔ a reviewer that hangs up the moment it has the whole bundle was served, not aborted (643.880041ms)
✔ send: a close after every byte left is served; one that fails a queued write is fetch-aborted (42.698245ms)
✔ a bundle with no bytes for bodyIdle refuses body-timeout long before its deadline (509.689803ms)
✔ send: a close once the loop is past its last chunk is served, and a failed write there is fetch-aborted (44.317684ms)
✔ letGo waits for an answer to leave for at most its cap, then destroys the socket (300.525188ms)
✔ a record that fails and can't be undone refuses rollback-failed, which the host logs (435.268345ms)
✔ land: the squash runs the gate in its own worktree, then main moves and the broker records it (404.375596ms)
✔ land: an arbiter-resolved landing's squash says the arbiter resolved it through the broker (343.366638ms)
✔ land: refusals before the gate answer the role at once and leave main and no worktree (345.727696ms)
✔ land: a failed gate is refused gate-failed with its step, told to the role, and main stays (340.572574ms)
✔ land: a record the broker refuses moves main back, and the landing is refused with the broker's code (340.635077ms)
✔ land: a second landing while one runs refuses land-busy, and close() stops the gate and refuses host-closing (348.946561ms)
✔ the host boots the broker, binds a launch in process, and the notifier DMs a blocking decision exactly once (1144.037832ms)
✔ a notifier that dies takes the host down with exit 1, so the unit restarts the pair (382.066909ms)
✔ a second host for the same data root refuses with exit 3 while the first runs (236.063305ms)
✔ two bus starts racing for one data root leave one host running; the other refuses that a host already runs (704.286885ms)
✔ a host takes over a host.lock its dead host left, and puts back a live claim that replaced it meanwhile (2.024841ms)
✔ a host.lock that goes between the check and the read is tried again; one that can't be parsed still refuses as unreadable (1.181933ms)
✔ a live host's lock stays, however old the file is (0.840762ms)
✔ a host.lock left by a host killed with SIGKILL doesn't stop the next bus start (1000.010099ms)
✔ while a host boots, a reader sees no host: the human commands still ask for --business, stop and status find none (269.436553ms)
✔ a host started before host.lock existed, holding only host.json, still refuses a second start, which leaves no lock (32.921864ms)
✔ a broker fork that throws leaves no host.lock and no host.json (33.940543ms)
✔ a broker send that throws leaves no host.lock and no host.json (36.795352ms)
✔ a broker that misses its boot wait leaves no host.lock and no host.json (36.800153ms)
✔ a notifier fork that throws stops the broker and leaves no host.lock and no host.json (138.445363ms)
✔ a notifier send that throws stops the broker and leaves no host.lock and no host.json (126.03193ms)
✔ a boot reply with no reader for the business refuses with exit 3 naming it, before the notifier starts, and leaves no host.lock and no live broker (129.355666ms)
✔ a boot reply with no reader for the business refuses with exit 3 naming it, with no notifier, and leaves no host.lock and no live broker (135.461072ms)
✔ a host.json that can't be written stops both children and leaves no host.lock (253.303861ms)
✔ an error event on the broker before it replies fails the start at once, and leaves no host.lock and no live child (37.773672ms)
✔ an error event on the notifier before it replies fails the start at once, and leaves no host.lock and no live child (134.858231ms)
✔ an error event on the broker while the notifier starts fails the start once the notifier replies (292.5302ms)
✔ an error event on the broker once the host is up stops the host with exit 1, and leaves no host.lock and no host.json (261.433195ms)
✔ an error event on the notifier once the host is up stops the host with exit 1, and leaves no host.lock and no host.json (284.534782ms)
✔ a broker reply that misses the wait breaks the channel: the late reply answers nothing, later requests refuse, the host exits 1 (1155.687353ms)
✔ a broker reply with another request's id, or none, breaks the channel and the host exits 1 (260.93489ms)
✔ a broker reply with no request waiting breaks the channel and the host exits 1 (145.325054ms)
✔ a notifier that refuses stops the broker and the host refuses with exit 3 (212.970349ms)
✔ a notifier that refuses after the broker died still refuses with exit 3, without a send to the dead broker (213.249404ms)
✔ a notifier that dies before it replies, after the broker died, still refuses, without a send to the dead broker (140.61719ms)
✔ a close send that fails with EPIPE after the notifier refuses still gives the notifier's refusal, exit 3 (197.179025ms)
✔ a close send that fails with EPIPE after the notifier dies unanswered still gives the notifier's error (143.737876ms)
✔ close() whose stop and close sends fail with EPIPE still finishes, with exit 1 (221.032179ms)
✔ watchChildren reports a child that died before it was called, and one that dies later (25.050028ms)
✔ bus stop refuses to signal a live pid that is not a bus host (203.122101ms)
✔ bus start refuses with exit 3 and the code when the broker refuses to boot; bus status names the lock (337.380126ms)
✔ bus start refuses with exit 3 without a notifier config (135.025986ms)
✔ bus start runs until bus stop; status reports it while it runs (815.683961ms)
✔ bus-service.sh renders the unit and installs it into a given directory (60.526857ms)
✔ bus-service.sh pins the claude binary, refuses one the unit can't carry, and renders without one (73.869096ms)
✔ bus-service.sh refuses a directory pin and keeps a symlink default unresolved (15.028127ms)
✔ bus-service.sh refuses a default claude without the execute bit, or in a PATH dir with a space (8.917226ms)
✔ row 165 CLI: instance add puts an instance in the file, validate exits 0, and the log names the human (219.864869ms)
✔ row 165 CLI: instance set passes the vars file and --no-launch, and a vars-only change needs no restart (222.420601ms)
✔ row 165 CLI: a refusal names its code and why, exits 2 and keeps the bytes (50.090448ms)
✔ row 165 CLI: usage errors exit 4 before any bus call, and an agent session is refused (24.498139ms)
✔ row 165: scripts/mosaic sends business instances and instance to packages/cli, and validate still to packages/business (203.488814ms)
✔ REQ-LIMIT-2 a claude-code launch asks mayLaunch with launch.account.name (26.78091ms)
✔ REQ-LIMIT-2 each login set the account can name is accepted when the bus host holds exactly it (44.947091ms)
✔ REQ-LIMIT-2 account-credential-missing: the named set is not a login set, or the bus host lacks part of it (44.501108ms)
✔ REQ-LIMIT-2 account-credential-ambiguous: the bus host sets a credential variable outside the named set, ANTHROPIC_BASE_URL included (30.502337ms)
✔ REQ-LIMIT-2 account-is-credential: launch.account.name is the value of a credential variable the bus host sets (21.508183ms)
✔ REQ-LIMIT-2 a pi launch carries account null and skips both account checks (20.65585ms)
✔ REQ-LIMIT-2 with no launch.account a claude-code launch carries null and the environment is not checked for ambiguity (11.603478ms)
✔ REQ-LIMIT-4 the launcher asks mayLaunch with the instance, provider, model and account (row 126), and goes on to authorization on an allow (28.730842ms)
✔ REQ-LIMIT-3 REQ-LIMIT-4 a launch whose subject is at its limit is refused limit-reached and logged with the check, limit and count (18.166213ms)
✔ REQ-LIMIT-4 fail closed: a refused op, a check failure or a malformed answer refuses limit-check-failed, naming the check and the code (114.717685ms)
✔ REQ-LIMIT-4 the launcher reaches token-limit only through the bus: launcher.mjs imports no limit check (0.903474ms)
✔ each launch rereads limits and model selections; lowered limits keep sessions and log exact hit (57.984988ms)
✔ row 65 a passing primary is the route: its provider and model reach mayLaunch, and a stated fallback is never used (29.281229ms)
✔ row 65 a failed primary preflight takes the stated fallback, and the limit checks see the fallback route (23.023617ms)
✔ row 65 a failed primary with no fallback refuses with the preflight code and never asks mayLaunch (15.31745ms)
✔ row 65 a fallback whose own preflight fails refuses with its code and names both reasons (12.219229ms)
✔ row 65 the fallback route spends launch.account when its module counts the account (17.918522ms)
✔ row 65 a missing harness, model or thinking var refuses no-route naming it, and the system execution keys never fill it (31.633987ms)
✔ row 65 a harness that is not switched on refuses harness-off, primary or fallback, and an off primary never falls back (44.909107ms)
✔ row 65 the loader refusing a module refuses harness-refused and names it; a module that throws refuses harness-failed (34.472303ms)
✔ row 65 a thinking level the route harness does not apply refuses thinking-unsupported, on the primary or a fallback taken (69.972201ms)
✔ bus start --pm: the launcher instance (the chief of staff) runs under its own PID namespace, registered, with a manifest (1451.316062ms)
usage: {"event":"usage-unknown","business":"acme","run":"re6ae5830162e","instance":"chief","turn":"98cb2c83-4fbe-42d8-9f7a-7e481151e829","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r2a6743c1693f","instance":"coder","turn":"8ae5041b-aa44-44f7-991f-85dd1a1a4285","source":"@none","code":"baseline-unknown"}
✔ the chief of staff launches a coder through its launch tool; the coder answers; refusals name their code (865.329447ms)
usage: {"event":"usage-unknown","business":"acme","run":"r7c6cc1baae9b","instance":"chief","turn":"82c33547-5a91-464d-ae60-500abb42fe78","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r36eb7cf094b2","instance":"pm","turn":"644ae13b-31e3-4b01-94d5-b55d40a1988c","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r7c6cc1baae9b","instance":"chief","turn":"0ce534d7-930f-4428-8c51-750cacd6a41c","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r7c6cc1baae9b","instance":"chief","turn":"d8e3e171-ab7f-4abf-abc2-5b35e85e3165","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r36eb7cf094b2","instance":"pm","turn":"bcf652d5-7d0c-430d-91fa-3766b1182ad8","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r36eb7cf094b2","instance":"pm","turn":"0903c142-4e25-47b1-b40c-27ff2a74b884","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r7c6cc1baae9b","instance":"chief","turn":"e37a66f3-f7ba-4231-9bb4-dda26f82d1ca","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r36eb7cf094b2","instance":"pm","turn":"4789f3a1-3d4c-429a-9197-4682d3e03ef5","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r7c6cc1baae9b","instance":"chief","turn":"345870b3-00ff-4191-a9c6-4baca556ec0c","source":"@none","code":"baseline-unknown"}
✔ a launched pm's request_staff reaches the chief of staff; no other session has the tool and the launch limits still hold (1067.77735ms)
✔ each launch makes sure the workspace has a clone of the project: cloned, fast-forwarded, or fetched and logged (3815.683265ms)
✔ a role the broker says holds a merge decision gets land_work, its prompt line, its landing and the git socket; without it, none (546.16465ms)
✔ REQ-LAUNCH-1 the tick starts a listed merge decision holder once, with land_work, the Landing line and its landing turn; an unlisted holder starts nothing (340.69499ms)
usage: {"event":"usage-unknown","business":"acme","run":"r82a7a694ee1f","instance":"coder","turn":"ca81d45a-ed00-444d-a14e-1c1ba56e3cbb","source":"@none","code":"baseline-unknown"}
✔ a holder running since before the resolution is stopped once the broker releases it, land-relaunch, and the next tick starts it: landing turn first, then the message that waited (592.578644ms)
✔ a landing run that ended isn't started again for its decision, a spent decision starts nothing, and a new decision starts the holder (484.526763ms)
✔ after a restart, the new host's first tick starts a holder whose decision is still held (683.596898ms)
✔ a refused landing start is logged once per code with its decision, isn't a start, and the next tick with room starts the holder (498.442772ms)
✔ a project clone that can't be made refuses checkout-failed, from --chief and from a session, with no run and no clone left (164.765596ms)
✔ a runner that stops at once ends its launch with the runner's reason (247.995725ms)
✔ a Claude Code launch with no session credential refuses before it starts (137.984519ms)
✔ a pi launch needs no session credential: pi logs in through its own ENV_ALLOW variables (243.336096ms)
usage: {"event":"plugin-refused","kind":"harness","name":"down","code":"missing","message":"/mnt/storage/src/mosaic-stack-rocko-178/packages/plugins/harness/down is missing","kept":[]}
plugins: {"event":"plugin-refused","kind":"harness","name":"down","code":"missing","message":"/mnt/storage/src/mosaic-stack-rocko-178/packages/plugins/harness/down is missing","kept":[]}
✔ row 65 a failed primary preflight launches the fallback on a fixture harness module, recording the route and the reason (266.516435ms)
✔ a runner killed from the host ends its launch as killed, not failed (293.360085ms)
✔ a missing or failing bwrap refuses sandbox-unavailable before anything starts; there is no unconfined fallback (141.612639ms)
✔ row 64 a role skill that isn't enabled refuses skill-not-enabled, and one that isn't a plain directory skill-refused, before anything starts (144.13259ms)
✔ row 64 on Claude Code a role skill that isn't enabled refuses the same way, before anything starts (139.639381ms)
usage: {"event":"usage-unknown","business":"acme","run":"r7f0b5c101080","instance":"chief","turn":"93000429-a0af-4854-91b7-136d5683a7f8","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r62db7dc96bb9","instance":"coder","turn":"1e5b9001-5998-4644-ad70-02f8c28bfa5c","source":"@none","code":"baseline-unknown"}
✔ a real Claude Code coder's shell can't read a stand-in secret or anything else its sandbox doesn't hold (1467.710895ms)
✔ a host that died hard leaves its sessions to the next host, which kills them and ends their runs so the role can be launched again (786.735484ms)
✔ a host that died hard leaves its sessions to the next host, which finds one already exited (23) and ends their runs so the role can be launched again (3688.55999ms)
✔ a malformed sessions.json refuses the host start with exit 3 and stays as it was (138.159388ms)
✔ an over-long launch request is refused at once, not at the 10 s idle timeout (131.853325ms)
✔ a launch client that never closes its side doesn't hold the host's close (132.005188ms)
✔ a runner that ignores SIGTERM is killed when the host closes (1259.952115ms)
✔ REQ-LIMIT-3 REQ-LIMIT-4 formatTokens shows each limit counted, not enforced, or check failed, in reset path words (1.324161ms)
✔ REQ-LIMIT-2 row 126: any daily quota row adds one line after the list saying the count is this stack's own use (0.492446ms)
✔ REQ-LIMIT-3 the view prints the limits then the token lines; --json keeps tokens in the result (4.216788ms)
✔ REQ-LIMIT-3 reset reads the revision, then sends limits.set with one path per limit, a word per level (1.114669ms)
✔ REQ-LIMIT-3 reset with no path, or with --file, is refused before the bus (3.329355ms)
✔ migration needs no bus/config, prints whole converted document, is idempotent and never writes (57.301827ms)
✔ limits view/set use the human bus transport and pass only the explicit patch (2.552167ms)
✔ limits set passes an assistants patch through unchanged, and the usage names the key (1.736635ms)
✔ post-119 example and Jason fixture migrate with checked inherited models; wrong or absent config refuses (308.933018ms)
✔ plugins view/set use the human bus transport and pass only the explicit patch (vikunja:19/305) (3.495441ms)
✔ inbox --read marks one message by the id inbox prints, delivered or not, and leaves the rest unread (52.328267ms)
✔ inbox --read refuses an id that is not the human's, prints its code, still marks the rest and exits 2 (34.424842ms)
✔ inbox --read refuses an ambiguous or short prefix, names every match, reads nothing for them and marks the rest (1.518061ms)
✔ inbox --read-all marks every message the inbox walk lists, delivered or not, and inbox then shows 0 unread (1471.552089ms)
✔ inbox --read stops on an error that is not about one message (0.864438ms)
✔ inbox --read and --read-all refuse bad combinations; other verbs refuse them (21.334649ms)
✔ the human names an instance, lists it, and talk's message.send reaches it by that name (34.138634ms)
✔ name and names refuse the wrong number of arguments with the usage (21.73329ms)
✔ zoned uses the IANA zone across DST (23.66694ms)
✔ each open blocking decision is DM'd once, across polls and a restart; non-blocking ones are not (35.115927ms)
✔ two blocking decisions get two DMs with different nonces (21.213732ms)
✔ the digest nonce differs per business and per day and fits Discord's 25 characters (0.401006ms)
✔ a failed DM is journaled, backs off, and is retried until it lands (25.703567ms)
✔ five definite refusals stop a DM: one gave-up line, one log line, and a restart keeps the count (35.884749ms)
✔ 429s, 5xx-style unknowns and refusals without a status never count toward the limit (29.500027ms)
✔ a crash between the fifth refusal and its gave-up line: the next poll appends it and sends nothing (21.916879ms)
✔ polled every POLL_MS against a permanent 403, a DM is sent at 0, 30, 60, 90 and 120 min and gives up only then (200.698467ms)
✔ a restart after the second refusal does not send before that refusal's 30 min are up (147.549733ms)
✔ the digest goes at 08:00 Chicago once a day, with blocking ones marked as DM'd (49.139257ms)
✔ a late start with no digest for the day sends one at once; an empty inbox gets one line (62.980895ms)
✔ an inbox read failure is logged and the next poll retries (1.776835ms)
✔ no Discord id reaches the journal or the log (41.943681ms)
✔ the journal: a torn tail is copied out and truncated, so an append after it reopens cleanly (10.773613ms)
✔ the journal: a crash between the copy and the truncate leaves a tail the next open repairs (11.535396ms)
✔ the journal: a whole file that is one torn line truncates to empty (7.530661ms)
✔ the journal: a malformed complete line refuses and leaves the file and any torn tail alone (1.468532ms)
✔ the journal: a loose file mode, a loose directory or a symlinked journal refuses (1.300792ms)
✔ the journal: a line with a wrong type refuses with exit 3 and names the field (4.361663ms)
✔ the journal: a symlinked directory refuses and says it is a link (0.763855ms)
✔ the journal: a dangling directory link, a parent that is a file and a journal that is a directory each refuse with exit 3 (0.972789ms)
✔ the journal: an append after the file was swapped for a symlink refuses and writes nothing through it (0.73977ms)
✔ the journal: a directory it cannot write or create refuses with exit 3 and names the path (0.771846ms)
✔ digest content stays within Discord's 2000 characters (0.487343ms)
✔ runLoop never overlaps ticks and stops after the one in flight (114.409385ms)
✔ a human command refused with human-required prints the other-tree line and exits 3 (3.504926ms)
✔ a --tester command refused with tester-required prints the other-tree line and exits 3 (1.43922ms)
✔ other exit-3 refusals print no other-tree line (4.52402ms)
✔ the line names no path (0.237967ms)
✔ threads project resolution explains a business repository and falls back if its file cannot load (65.508778ms)
✔ project create prints the repository-name refusal and exits two without writing (29.569127ms)
✔ project init/check use declared business repo and never contact the bus (228.89764ms)
✔ project commands validate args and refuse undeclared projects and agent init (86.056769ms)
✔ project create and list validate their arguments before any bus call (89.260349ms)
✔ project dispatcher reaches the human CLI (210.073464ms)
✔ shared copy: npm ci, pushed reachability, readonly/existing copy, webui delegation (1098.398387ms)
✔ wrapper preserves Q81 lease and marker, rejects foreign/symlink; check uses effective roots (767.2021ms)
✔ bus unit copy/pm pins, install refuses all busy states and failed reads, drop-in refusal (693.257107ms)
✔ missing package manifest refuses before npm; npm failure removes the unpublished temp copy (502.463762ms)
✔ real pinned install loads bus and Console; same-copy mosaic talk passes proof, other trees refuse (3882.441337ms)
✔ bus status reads effective copy/pm/drop-ins without writes; check refuses an overridden ExecStart (650.53677ms)
✔ drop-ins in either override field and malformed metadata refuse; uninstall also checks --no-reload (284.981421ms)
✔ install never shows a template name on systemd 261; template drop-ins still refuse (278.846568ms)
✔ unlisted instance drop-in is found in another unit lookup directory; path discovery errors refuse (273.614911ms)
✔ inherited mosaic-.service.d override refuses without changing the unit (277.846978ms)
✔ inherited service.d override refuses without changing the unit (295.316994ms)
✔ inherited drop-in unreadable directory or file refuses before writing (260.245759ms)
✔ T1 adopt writes the pointer from the units once; a failing check refuses check-failed (1474.917445ms)
✔ T2 activate installs exactly the pointer's pm, claude and Console flags (1747.373448ms)
▶ T3 a hand-edited unit refuses unit-flags-drift, names the difference and changes nothing
  ✔ pm (871.868194ms)
  ✔ claude (791.81025ms)
  ✔ console (730.495743ms)
✔ T3 a hand-edited unit refuses unit-flags-drift, names the difference and changes nothing (2395.145689ms)
▶ T4 each item 3 refusal changes nothing and logs one line with its code
  ✔ copy-missing (697.276836ms)
  ✔ copy-unreachable (659.934718ms)
  ✔ copy-schema-unknown (684.735809ms)
  ✔ already-active (634.780749ms)
  ✔ instance-busy (another business active) (627.992182ms)
  ✔ instance-busy (this business activating) (620.658088ms)
  ✔ drop-in-override (650.562878ms)
  ✔ no-pointer (322.459682ms)
✔ T4 each item 3 refusal changes nothing and logs one line with its code (4899.936279ms)
✔ T5 an activate with no migration stops the Console then the bus, installs, starts the bus then the Console (1591.933034ms)
✔ T6 a migration without --migrate rolls back with migration-needs-flag (1538.857156ms)
✔ T7 --migrate takes the backup before the migration runs (1518.520469ms)
✔ T8 a failed migration restores the backup and keeps the store it replaced (1551.613808ms)
✔ T9 a store outside the target's chain rolls back with store-newer (1419.709818ms)
▶ T10 each gate check failing alone rolls back and puts every file back
  ✔ units-inactive (4738.863868ms)
  ✔ bus-status (5114.482884ms)
  ✔ healthz (4304.934812ms)
  ✔ copy-check (4365.299482ms)
  ✔ pm-launch-unconfirmed (5724.973137ms)
✔ T10 each gate check failing alone rolls back and puts every file back (24249.514047ms)
▶ T10b the PM check needs exactly one launched line for a live run of this start
  ✔ no launched line (4455.154475ms)
  ✔ two lines (4518.917392ms)
  ✔ a changed word (4449.450037ms)
  ✔ run not in launches list (4622.012172ms)
  ✔ run started before the start (4607.843589ms)
  ✔ a human launch with no journal line (4502.714058ms)
  ✔ run for another business (5906.815051ms)
  ✔ run not live (4645.502597ms)
  ✔ one well-formed line and a live run pass (1490.636109ms)
✔ T10b the PM check needs exactly one launched line for a live run of this start (39200.540823ms)
✔ T11 a rollback whose gate fails too: rollback-failed, exit 2, both units stopped, no retry (8840.608869ms)
✔ T12 rollback takes the previous copy from the log (2512.004197ms)
✔ T13 a rollback across a migration needs --restore, then restores and goes live (3177.818146ms)
✔ T14 events written after the backup need --events N with the right N (3362.328803ms)
✔ T14b a rollback --restore whose gate fails stops both units and skips item 6 (6165.88373ms)
✔ T14c a rollback without --restore whose gate fails runs item 6 (6321.544732ms)
✔ T15 --plan names the migration and the restore and changes nothing (1874.122413ms)
✔ T26 --plan lists the commits an activate adds, after the migration lines (933.613196ms)
✔ T27 --plan names each migration step, or says the steps cannot be read (783.528888ms)
✔ T25 --plan refuses copies-unrelated when neither copy is an ancestor of the other (339.410875ms)
▶ T23 --plan reads the store without changing a file, stopped or held
  ✔ stopped store (436.95262ms)
  ✔ held store (408.307309ms)
✔ T23 --plan reads the store without changing a file, stopped or held (845.773828ms)
✔ T24 business-mismatch and capture-failed refuse with exit 3; a short id is a usage error (1059.322324ms)
✔ T16 status prints the pointer, the log, the check, bus status and drift, and writes nothing (640.435065ms)
✔ T19 a started line with no final line: status shows it and every changing verb refuses (883.654637ms)
✔ T20 a held run lock refuses run-locked without a log line; status and --plan still run (926.124499ms)
✔ T21 a bus unit with no MOSAIC_CLAUDE_BIN: adopt and activate refuse no-claude-pin (1085.198986ms)
▶ T22 the stopped store read: a -wal or -shm refuses store-unclean and is left alone; a clean read adds no file
  ✔ bus.sqlite-wal (1359.461794ms)
  ✔ bus.sqlite-shm (1399.417245ms)
  ✔ clean (1424.992757ms)
✔ T22 the stopped store read: a -wal or -shm refuses store-unclean and is left alone; a clean read adds no file (4184.466619ms)
▶ T28 a unit path, drop-in directory or drop-in file that cannot be read, or a relative unit path, refuses drop-in-override before any stop
  ✔ unit path (588.579476ms)
  ✔ drop-in directory (599.024441ms)
  ✔ drop-in file (586.644217ms)
  ✔ relative unit path (617.184468ms)
✔ T28 a unit path, drop-in directory or drop-in file that cannot be read, or a relative unit path, refuses drop-in-override before any stop (2392.164471ms)
▶ T29 a bus that stops during the held-store read: --plan says so, rollback --restore refuses restore-mismatch
  ✔ --plan (599.898775ms)
  ✔ rollback --restore (1822.319149ms)
✔ T29 a bus that stops during the held-store read: --plan says so, rollback --restore refuses restore-mismatch (2422.751476ms)
✔ T30 a copy check that agrees on another copy fails the gate with copy-check and rolls back (4147.938302ms)
▶ T31 a quoted or escaped Environment= or ExecStart= is unreadable: refusals by name, nothing read from /a
  ✔ Environment double quotes (1291.862153ms)
  ✔ Environment single quotes (1266.500763ms)
  ✔ Environment backslash (1259.117359ms)
  ✔ ExecStart quotes (1307.557051ms)
  ✔ Console ExecStart quotes (1263.482832ms)
✔ T31 a quoted or escaped Environment= or ExecStart= is unreadable: refusals by name, nothing read from /a (6389.455181ms)
▶ T32 item 1: --pm-after-harness refuses a target without harness or a file that sets it, and passes without the PM
  ✔ usage: not with --pm, and only for activate (476.38594ms)
  ✔ pm-after-harness-target (724.117317ms)
  ✔ pm-after-harness-set (1028.295366ms)
  ✔ a pointer without pm (666.596981ms)
  ✔ phase 1 installs without --pm, skips pm_launched and writes the pending pointer (1467.277127ms)
✔ T32 item 1: --pm-after-harness refuses a target without harness or a file that sets it, and passes without the PM (4363.528307ms)
▶ T33 item 2: read_pointer takes pmPending only as true beside pm false; a pending pointer refuses activate and shows in status
  ✔ read_pointer (1049.520505ms)
  ✔ pm-pending, already-adopted and the status line (2088.145862ms)
✔ T33 item 2: read_pointer takes pmPending only as true beside pm false; a pending pointer refuses activate and shows in status (3138.149569ms)
▶ T34 item 3: pm-on refuses before any stop, passes with the PM, or puts the phase-1 bus back
  ✔ not-pending (533.642695ms)
  ✔ harness-unset (1749.206665ms)
  ✔ a launched line passes: the bus alone restarts with --pm (2573.680236ms)
  ✔ a failed pm_launched puts the phase-1 unit back: exit 1, still pending (5246.96624ms)
  ✔ a failed phase-1 gate too: pm-on-down, exit 2, both units stopped (8720.661949ms)
✔ T34 item 3: pm-on refuses before any stop, passes with the PM, or puts the phase-1 bus back (18825.047119ms)
▶ T35 item 4: rollback from pending installs --pm; a plug-in kind the target lacks refuses by name
  ✔ rollback from pending gates pm_launched on the old copy (2549.471108ms)
  ✔ after the harness set, rollback refuses until harness is gone (3737.146603ms)
  ✔ activate names every kind the target lacks (918.820799ms)
✔ T35 item 4: rollback from pending installs --pm; a plug-in kind the target lacks refuses by name (7206.035359ms)
✔ T36 --plan shows the pending state and lists the row 169 refusals it can see (2489.934235ms)
✔ T37 tester-port activate writes pointer and unit, replaces the port, and supported re-pin and rollback keep it (4814.923349ms)
✔ T38 tester-port rollback to unsupported target drops flag in plan, steps, unit and pointer (2604.973025ms)
✔ T39 tester-port explicit unsupported target reads target copy and refuses before stopping either unit (582.460523ms)
✔ T40 tester-port invalid values refuse exit 4 before stopping, including tunnel and main ports (564.076318ms)
✔ T41 tester-port inherited activate to unsupported target drops it and retains drift refusal and failed-run recovery (7952.239329ms)
✔ T42 tester-port pm-on preserves pointer Console flags (2381.208363ms)
▶ T43 row 180 item 1: runs in one second get distinct activation ids, rollback sets and post-attempt stores
  ✔ activate then pm-on on the same copy (2379.114066ms)
  ✔ two restores: an old-form id and a suffixed one, each with its own post-attempt store (4910.771352ms)
  ✔ adopt takes the next free id too (458.556305ms)
  ✔ all ten names taken: capture-failed before anything stops, and the log line names no activation (682.049001ms)
  ✔ --restore takes an id with or without -1 to -9 and nothing else (701.0352ms)
✔ T43 row 180 item 1: runs in one second get distinct activation ids, rollback sets and post-attempt stores (9132.507272ms)
▶ T44 row 180 item 2: --plan lists tester-port-invalid and tester-port-target with its other refusals
  ✔ a bad tester port (1491.914619ms)
  ✔ without --plan a bad port is usage before any pointer read (263.450781ms)
  ✔ a target Console without --tester-port (723.074746ms)
✔ T44 row 180 item 2: --plan lists tester-port-invalid and tester-port-target with its other refusals (2479.178607ms)
✔ T178 proxy proof header and path survive re-pin and rollback unchanged in the pointer and unit (2453.274053ms)
✔ T178 proxy proof pair survives a tester-port replacement as well (1435.6189ms)
✔ row 175: each read with --tester inside an agent prints the human form's output, and refuses without --tester (171.828313ms)
✔ row 175: --tester on limits set, reset and migrate and every write form outside W1-W7 exits 4 before any call (64.732609ms)
✔ row 175: scripts/mosaic sends plugins, name, names, follow, threads and thread to packages/cli, not the seat CLI (929.61817ms)
✔ row 175 W7: project init --tester makes the repository under the project area, committed by tester (421.076843ms)
✔ row 175 W7: a repo outside the project area refuses tester-project-area before any bus call, and nothing changes (176.154456ms)
✔ row 175 W7: the tester bind refuses tester-off and tester-required before any directory (1079.670097ms)
✔ talk --tester sends a message as tester from inside an agent and prints the role's reply (63.274317ms)
✔ talk --tester sends text with no prefix, and follow --tester reads (32.940077ms)
✔ the read verbs take --tester inside an agent, and the trail shows the tester as the sender (91.909306ms)
✔ decide, stop and launches take no --tester (38.558657ms)
✔ the tester transport runs tester-cli.mjs, and the tester's refusals exit 3 (0.83831ms)
✔ with no stand-in transport, --tester runs tester-cli.mjs on the data root's socket (1.154255ms)
✔ talk --tester to a stopped role exits 3, prints no sent line and stores nothing (51.540909ms)
✔ only a --tester send refused tester-target-not-running is rewritten (33.468349ms)
✔ REQ-CLI-1 project create and list work on the bus record and refuse what the broker refuses (106.252465ms)
✔ REQ-CLI-1 thread create, list, open and read (138.399945ms)
✔ REQ-CLI-1 thread read pages through every message, oldest first (530.627412ms)
✔ REQ-CLI-1 the business needs exactly one threaded instance to create a thread (53.670496ms)
✔ REQ-CLI-1 a thread is its id, an id prefix of 8 characters or more, or <project>/<name>; anything else is refused before any write (67.470613ms)
✔ REQ-CLI-2 REQ-THREAD-2 talk --thread sends to the thread and prints only the reply to it; the rest stays unread (43.185432ms)
✔ REQ-CLI-2 talk --thread --wait gives up with exit 1 and leaves the thread's messages unread (1041.817556ms)
✔ REQ-CLI-2 talk --thread --follow and follow --thread print and read only that thread's messages (95.70336ms)
✔ REQ-CLI-2 plain talk to the assistant asks for a thread and exits 2 without a write (30.116226ms)
✔ REQ-USER-2 a thread of another user is refused for talk, open, read and follow, and nothing is written (40.607029ms)
✔ REQ-CLI-1 the new verbs refuse inside an agent run, as every human verb does (13.253253ms)
usage: {"event":"usage-unknown","business":"acme","run":"rd7ab1ddaf8cb","instance":"assistant","turn":"b4bf2acd-7b70-4dc6-ba65-68ea11e6dca0","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r4d819c91be5d","instance":"assistant","turn":"a7139c5b-c009-4375-b19e-fd10496af5d9","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r294085cc08e8","instance":"assistant","turn":"ef85c0c9-cfc5-4959-8779-bf58f9d30f2d","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"redabefaf61b5","instance":"assistant","turn":"f9819fed-d378-4c19-9c3a-e2e28e1097c3","source":"@none","code":"baseline-unknown"}
✔ REQ-THREAD-2 REQ-LAUNCH-1 REQ-MSG-1 a thread message starts the thread's session, which stops when idle; its history survives the stop and a restart and no other thread sees it (5763.199279ms)
usage: {"event":"usage-unknown","business":"acme","run":"rad74e264aab8","instance":"assistant","turn":"8d1805f4-9dfc-4d9d-9c21-725a4dbf6d0d","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"r34cb48b45943","instance":"assistant","turn":"5ed34aea-b807-4426-b76e-6997151599fc","source":"@none","code":"baseline-unknown"}
✔ REQ-THREAD-2 REQ-LAUNCH-1 a running thread counts against its model's limit and an idle-stopped one doesn't (1657.633641ms)
usage: {"event":"usage-unknown","business":"acme","run":"r2459027b67c9","instance":"assistant","turn":"f2b44a80-4ce4-4971-8299-4e93198e4aed","source":"@none","code":"baseline-unknown"}
✔ REQ-THREAD-2 REQ-LAUNCH-1 a turn longer than the idle time isn't stopped, and the thread idles out after it (4425.037412ms)
usage: {"event":"usage-unknown","business":"acme","run":"rba45f504a7ac","instance":"assistant","turn":"82f6d149-77e2-4951-b969-e0e8eb9fc869","source":"@none","code":"baseline-unknown"}
usage: {"event":"usage-unknown","business":"acme","run":"rba45f504a7ac","instance":"assistant","turn":"ad32718f-30df-45a5-bb67-c2a0f05fc987","source":"@none","code":"baseline-unknown"}
✔ REQ-THREAD-2 REQ-LAUNCH-1 a message sent after the idle snapshot keeps the session, which answers it (2855.004521ms)
✔ REQ-ROLE-2 REQ-LAUNCH-1 a thread held by a run the launcher didn't start gets no second session (51.139733ms)
usage: {"event":"usage-unknown","business":"acme","run":"r50188db9dff0","instance":"assistant","turn":"bec993d6-7251-437d-9051-55e29bae5748","source":"@none","code":"baseline-unknown"}
✔ REQ-LAUNCH-1 a thread run's session file has no landing, even when landHolders names its instance (283.16349ms)
✔ REQ-LAUNCH-1 a run that ends without taking its thread's message isn't started again for that message (154.553716ms)
✔ REQ-THREAD-2 REQ-LAUNCH-1 a thread id that isn't a safe segment, or a thread of a non-thread instance, launches nothing (186.282522ms)
✔ REQ-THREAD-1 REQ-LAUNCH-1 the chief can't launch the assistant outside a thread (52.722526ms)
limits: {"event":"limit-reached","business":"acme","check":"token-limit","export":"mayLaunch","subject":{"instance":"assistant","provider":"zai","model":"claude-sonnet-5-5","account":null},"path":["tokens","instances","assistant"],"limit":0,"count":0,"unknownTurns":0}
limits: {"event":"limit-reached","business":"acme","check":"token-limit","export":"mayLaunch","subject":{"instance":"assistant","provider":"zai","model":"claude-sonnet-5-5","account":null},"path":["tokens","instances","assistant"],"limit":0,"count":0,"unknownTurns":0}
✔ REQ-LIMIT-3 REQ-LIMIT-4 a thread start whose instance is at its token limit is refused, logged once with the limit and the count (168.142469ms)
usage: {"event":"usage-unknown","business":"acme","run":"r861692ac6abb","instance":"assistant","turn":"1365c50d-43b6-4776-b56f-aca2fa0b5634","source":"@none","code":"baseline-unknown"}
limits: {"event":"limit-reached","business":"acme","check":"token-limit","export":"mustStop","subject":{"instance":"assistant","provider":"zai","model":"claude-sonnet-5-5","account":null},"path":["tokens","instances","assistant"],"limit":0,"count":0,"unknownTurns":1}
✔ REQ-LIMIT-4 a thread run at its token limit stops at the end of its turn: the reply lands, and the end record names the limit (278.59655ms)
task.close {} answered: invalid-request; fake saw 18 requests, first GET /info 200, GET /projects/1 200, GET /projects/1/views 200
task.close on a missing task answered: task-not-found; it made GET /tasks/999 404
✔ bootConfig trackers reach the S3 adapter in the real broker child, which goes ready against a fake Vikunja (425.405845ms)
✔ the transport writes {business, verb, args} to the child and reads its JSON (48.806688ms)
✔ a bus code on stderr becomes the exit code; garbage and timeouts are outcome problems (2248.423673ms)
✔ busExit and refuseInsideAgent (0.587806ms)
✔ REQ-LIMIT-4 mosaic usage shows each token kind and unknown turns beside every total (122.909294ms)
✔ usage is dispatched by scripts/mosaic and the shared read adapter (154.373232ms)
✔ talk sends a REQUEST, prints and reads only the reply to it, and leaves other messages unread (72.546975ms)
✔ talk --wait 0 only sends; no reply within --wait exits 1 and says where it will show (1040.323828ms)
✔ a late reply the Console has listed stays unread for mosaic inbox, and the next talk skips it (2048.320272ms)
✔ talk says when the role isn't running, whether launch.by was told, and how to start launch.by (34.852528ms)
✔ talk to a role launch.by can't launch says it isn't in the launch block (14.580476ms)
✔ talk to a stopped role prints the waiting line once, skips a non-reply and prints the reply (47.734722ms)
✔ talk, stop and launches off/on refuse inside an agent run; bad arguments are usage errors (24.404945ms)
✔ launches off and on go to the broker and change the business's launch state (17.931535ms)
✔ launches list reads sessions.json and marks a stale entry; stop reports it and refuses a non-runner (16.257554ms)
ℹ tests 408
ℹ suites 0
ℹ pass 408
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 213076.086746