propose_skill took no strict schema, so a runbook sent under another key
was dropped and the saved one proposed again; the confirm still said
"Saved". Now an unknown key is an error, a change that changes nothing is
refused, and both the bot and the confirm outcome are told which fields
change and whether the runbook is among them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Each computer's Chrome profile now lives on the data volume
(DATA_DIR/computer-browsers/<id>, mounted at ~/.cobble-browser) and
Chrome is pointed at it with --user-data-dir. The sandbox image resets
its own profile's preferences to session-only cookies on every boot,
which signed every site out whenever Chrome exited. New profiles are
seeded with the image's preferences minus that rule, stale Singleton
links are removed before each provision, and existing computers are
recreated once (skipped while a job runs on them).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Bots get list_hooks, pause_hook and delete_hook, which act directly since
they only take away. A new hook, or a new name, brief or filter, or
switching a paused one back on, goes through propose_hook: it stores a
hook_proposals row and puts a [hook-proposal:ID] card in the reply, which
the user it was proposed to confirms or rejects. The card is read from
what Confirm saves, with the bot's reason labelled as its own; confirming
checks everything again and saves through the Hooks modal's helpers. A
new hook's URL is shown on the card once, to the user only, never in the
outcome the bot is told. What calls a hook, its computers and its signing
are never proposed. Only chat replies propose, and a hook call's own run
gets no hook tools, so a caller's payload can't reach them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A bot hook (bot_hooks) is a secret URL, POST /api/hooks/<token>, whose
calls queue the hook's brief plus the raw body as a task, per user and
bot like a routine. Only the token's hash is kept (shown once, rotatable);
an optional HMAC-SHA256 signing secret is stored encrypted. Calls are
rate limited per address and per hook, recorded in bot_hook_events
(pruned daily), deduplicated by delivery id, and a call while the last
run still waits joins its brief instead of queueing another. The payload
is fenced data and HookRunNote tells the bot never to take instructions
from it; failed runs retry like routines.
A hook can instead watch a computer's browser notifications: the app
installs a small watcher that owns org.freedesktop.Notifications on the
sandbox's session bus, so Chrome hands it every notification (it only
looks for that service at start-up, so Chrome is restarted once, never
mid-task), and it posts each one to its hooks, optionally filtered.
Computers reach the app at COMPUTER_APP_URL or the app container's name.
Users make hooks in the conversation's Hooks modal and manage them all
on /admin/routines; bots never create or change hooks.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A bot's instructions and memory for one workspace were only reachable
from the workspace's settings. The editor is now shared and also opens
from the conversation menu in that workspace ("Instructions here",
loaded on open through api/admin/workspaces/workspace-bot) and from the
workspace rows on the bot's settings page.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The conversation header menu gets a Copy chat row that puts the active
session on the clipboard as one markdown document: a summary, then each
message under its author and time, with quotes, attachments and
reactions. Bot and workspace refs read as names, message headings drop
below the message headings and code blocks stay untouched. The clipboard
write starts on the click with a pending ClipboardItem so Safari allows
it after the fetch.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The wrap-up turn now gets the run's actual tool calls, each bot message
keeps them as bot_messages.tool_log, and history follows recorded bot
messages with an [App record] line. The coordinator also gets update_bot
in workspaces, so it can fill a bot's core prompt from there.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
delegate_task takes a task_id to hand over a task already on a list: it
moves to the worker (a workspace's task stays on its list), goes to
doing, and the job is linked back so its end is noted on the task.
hand_over_schedule moves a bot's own routine to another bot, keeping its
workspace. Users get Hand over in the Tasks UI and both routine lists.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Settings > General gets an App icon (super admin): one of the seven
stones, or an uploaded image with the picked stone as a top-left corner
mark (inside a round launcher's crop, clear of the browser's badge),
previewed in both launcher crops before saving.
Uploads are decoded and cropped to a stored square (no SVG); the set is
drawn into DATA_DIR/install-icon/<version>/ and served publicly and
immutable. The manifest, head icon links and push notifications point at
it; the ember stone keeps the stock icons. The generator shares the icon
renderer.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Settings > General: an App name (super admin) for the installed app.
The manifest's name/short_name, the iOS head tags and the loading
page's title use it, so two installs on one phone can be told apart.
- Notifications get a real badge (white stone silhouette): Android masks
the badge by alpha, so the opaque app icon showed as a square.
- The loading and offline pages load the app's font.
- The font is now Archivo (variable, one request for every weight),
matching the marketing site.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The longest runbook a skill may save and how much of it load_skill hands
back are now set on Settings > General (skill_body_max_chars,
skill_body_shown_chars), defaulting to AppData.SkillBodyMaxChars and
SkillBodyShownChars. Saving, confirming and proposing a skill check the
install's limit; a skill saved before it was lowered can still be toggled.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The bare "logs" pattern ignored every folder named logs, so
api/admin/logs/clear-logs.ts never reached the repository or the
production build. Anchor it to the root and add the route.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
startDataBackupRun already holds the backup run lock, but the archive
(createDataBackup) and restoreDataBackup took it again, so every run
failed a few seconds in with "Another backup or restore is already
running". The writer and the restore now take no lock of their own.
The run's notice (archiving, restoring, or why it failed) moves to its
own line under the actions row so the buttons don't shift, and archiving
reads in gray rather than the warning colour.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A bot's powers came from bots.kind alone, so only the coordinator could
hand work to other bots. bot_types now works like user_types: a row with
no workspace holds everywhere (the coordinator), one with a workspace
only there (a workspace admin), and no row is a worker. bots.kind stays
the legacy marker of the one coordinator for who sees it.
The matrices in src/dict/bot-types-dict.ts say what each type may do,
whom it may direct and who grants it, read by one generic interpreter
(rbac/type-assignment.ts) that userTypeAssignment now wraps too. Every
run reads its powers once (grabBotRunPowers) and the chat tools, task
runs and task tools gate on them instead of kind; a run's powers are its
type's and its user's together.
In the workspace it leads, an admin hands tasks to the workers there
(results relayed back to it), sets their instructions, adds tasks to
their lists and keeps the workspace's context when its user may. Super
admins and admins grant it from the workspace's settings or the chat
header's menu, and the coordinator with set_bot_type. Admins are marked
with a filled green shield beside their name across the workspace, kept
live by pushBotUpdated's admin_workspace_ids.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A data archive or restore ran inside its request, which sent nothing
until tar finished: Bun closes a silent socket after 10s (its default
idleTimeout in production) and a proxy cuts it at about 100s, so the
page showed a network error while the archive kept going, and a retry
said another backup was already running.
The routes now start the run (startDataBackupRun, refused while the
backup lock is held) and answer at once. The page follows it by polling
api/admin/backups/data-status: an archive reloads once written, a
restore once the status comes from a new process after its restart,
and a failure shows its error. A run in progress survives a reload.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Computers now mount only /home/gem/workspace from DATA_DIR/computers/<id>,
so only the workspace persists and goes into the data archive. An older
computer's whole home is moved aside to <id>-home-<ts> (kept, never
archived) and its container recreated once by sync-computers.
Each run works in bots/<bot>/users/<user> (also its cwd and HOME), with
bots/<bot> for work every user shares and workspaces/<id> shared by a
workspace's bots; the coordinator reaches the whole workspace. Computer
tools are gated by withComputerDirs: bash and code start there, file tool
paths and file: URLs outside the run's folders are refused, and so is sudo.
search_computers, import_file and script tools start there too.
The files browser shows members only the bots' folders they reach, their
own user folder in each and their workspaces', with names on the id folders.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Coordinator tools in its own chat (super admin only): list_extensions,
set_extension_access, delete_extension, and update_skill /
update_mcp_server / update_custom_tool for what can't change where
anything is sent (on/off, names, arguments, left-out tools)
- Adding a skill, MCP server or tool, or changing a skill's words, a
server's URL or headers or a tool's kind or config, goes through
propose_skill / propose_mcp_server / propose_custom_tool: stored in
extension_proposals with the bot's background check, shown as a
[proposal:ID] chip that opens a card with Confirm and Reject
- The card's facts are worked out by the app from what Confirm saves: the
host, every {{env.KEY}} and whether it's saved, the config; for a skill
its links, hosts, tool names, mentions of secrets and a diff against
the saved runbook
- Confirming decides once and checks again; a server or tool missing a key
is saved switched off, and the outcome is posted to the bot in chat
- Never from a bot: stdio servers, a URL whose host a template decides,
envs in a prompt tool, a key's value
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Activity page: "Delete all" empties the open tab (api/admin/logs/clear-logs,
super admin only, logged under system)
- Stop logging every WebSocket connect
- Log unexpected API errors (safeErrorMessage -> logUnexpectedError), every
failed bot job attempt and every job the queue lost
- Drop the cron jobs' console output; a failed model sync goes to the logs
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Both tables lay their heading and rows on one shared grid, so every
figure sits under its own heading (twui's Row wrapped the cells).
- The selected range reads as pressed: ink-filled in a hairline track.
- Each day's bar opens a small summary card on hover or tap.
- twui 0.1.7 places popovers before paint, so opening one no longer
makes the layout flicker.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A reply read the session once at the start, so a message sent while it
ran was never seen, and the follow-up reply queued for it then finished
silently: the message was saved before the reply, so the thread ended
with the bot. Those messages went unanswered.
The text loop now reads new user messages after every round of tools and
before an answer goes out, and the bot decides from what each one means
whether it changes the work (folded in, or the unsent answer redrafted)
or is a separate request, which it acknowledges at once with the new
send_message tool and does after the current work. Messages sent that
way are marked interim, so the thinking row stays up while the bot
works. A reply job records the newest user message it read, and a queued
reply only stays silent when a finished reply already read everything.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Every bot message gets an info button that opens its details: the
provider and model behind it, tokens and cost, the job that ran it, each
model used and every call in order. Each model call now records how it
was made in token_usages.settings (generation settings sent, whose
settings they were, pinned or automatic model, fallbacks), shown per
model in the modal.
Live replies now carry their token usage, from the job's push and the
live relay alike, so it shows without a reload.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A bot's or workspace's picture was only shared when its media row carried
the bot-image or workspace-image paradigm. A picture set before pictures
were copied into their own row points at the chat upload itself, so only
its uploader could load it and everyone else saw the default icon. A
picture is now shared because a bot or workspace points at it, whatever
its paradigm; profile photos still have to be the user's primary one.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The conversation pages rendered their messages on the server, so a page
the back button restored from the back-forward cache showed an old
thread. The server now renders everything else (rail, header, composer)
and the history (active session, latest messages, active task) loads
from api/admin/bots/conversation-history once the page is on screen,
with a loading line under the header, and again on a pageshow restore.
Messages sent or pushed while it loads are merged with it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Media rows can no longer be created through the CRUD API, updates only
take description fields on the caller's own row, and every media read
or delete goes through mediaFilePath, which refuses paths outside
DATA_DIR or inside the database folder.
- Uploads strip the file name and extension to safe characters and keep
only title, description and summary from the request's media object.
- Folder imports no longer follow symlinks (walkRegularFiles), and the
source path is quoted in the computer's shell.
- User and bot files are served with nosniff, and HTML, SVG, XML and
scripts are sandboxed by CSP.
- Sandbox cookies are kept to the computer's prefix, and websocket
upgrades on /ws and the computer proxy must come from the app's own
address (README and Help note nginx's Host header).
- user_types CRUD writes go through the assignment matrix; members only
read their own user record; read-only users can read CRUD again;
grabUsers builds valid SQL for array search terms and bad pages; an
admin media DELETE needs an id; read-only users can't delete media.
- Rate limits and the audit log use the connection's address, trusting
forwarded headers only from a proxy, and login limits are split per
address, per account from that address and per account overall.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A computer a bot names in a reply is now a button that opens it in the
shared computer viewer modal, the bot's own computer included (sunken and
checked, as before). Switching stays in the header's computers modal, so
the chip has one job. A computer the bot can no longer use stays a plain
name.
The bot's id never reached its prompt: `composeBotPrompt` only took its
instructions, the appendix, its memory and its workspace, and
`grabBotPrompt` never passed `bot.id` or `bot.name` on. A bot could only
learn its id by accident, from a tool description that named it, so
"which bot am I" was unanswerable in most contexts.
- Open the prompt with who the bot is (`composeIdentity`): its name,
whether it's the coordinator or a worker, its role, and its id, said
to be how tools and refs point back at it. `grabBotPrompt` now passes
the bot, which covers chat replies, tasks, routines and delegated
tasks, since both runners build their prompt there.
- Turn the notes' example refs into functions bound to the running bot:
`chatReplyNote`, `botFilesNote` and `coordinatorReplyNote` take its id,
so the folder-ref example no longer points at a hardcoded `owner_id: 5`
(another bot's folder) nor `[bot:3]`. ComputerFilesNote keeps the
computer it's given.
- Stop `create_task` from hiding the worker's own id in its own chat.
Tests: four new `composeBotPrompt` cases pin the identity block (worker,
coordinator, unnamed, no bot at all), and a new `harness-prompts.test.ts`
checks each note's own-folder and self refs point at exactly the running
bot, for three ids.
The trail is twui's Breadcrumbs reading Home, the sections and the page
off the address, rendered as the first thing in SettingsShell's content
column, so every page that is not a conversation gets one without a page
touching it.
The last crumb is named by the page's own title rather than the address:
a record reads Scout instead of its id, /admin/logs reads Activity, and
/admin/settings/mcps reads MCPs. That replaces the breadcrumb_title prop
the four settings record pages had to pass, so settings-breadcrumbs.tsx
goes.
hasBreadcrumbTrail drops the trail on the crew of bots, home, and on
anything outside /admin; the trailing slash and query string are ignored,
so a folder in the files browser keeps the trail of the page it is on.
Say it once in AGENTS.md: a className is twMerge(...) with one chunk per claim, so a later chunk beats an earlier one instead of depending on CSS source order.
The two limits were compiled in: a task run stopped at
AppData.HarnessMaxToolSteps (200) and a chat reply at
AppData.BotReplyMaxToolSteps (8), so an install that needed a longer task
had to rebuild.
Both are now `app_settings` rows (`harness_max_tool_steps` and
`bot_reply_max_tool_steps`), read at the two loop call sites through
grabHarnessMaxToolSteps and grabReplyMaxToolSteps, and edited by the super
admin on Settings > General beside the prompt appendix. Zero clears a
limit back to the code default, like the bot file quota.
readStepLimit keeps the saved number only when it is a whole number of at
least 1 and no more than AppData.BotStepLimitMax (1000), so a blank, a
zero or a typo falls back instead of stopping every run before its first
turn, and a huge number still ends in the wrap-up report.
runTextLoop keeps AppData.HarnessMaxToolSteps as its default, so it stays
free of database reads and its tests run as they did.
Sending a message while scrolled up left it out of sight: the thread only
ever anchored its bottom, so nothing brought the view down. The column now
follows new messages while it sits at the bottom, stays put while it's
scrolled up, and always brings a message the user just sent into view.
While the thread sits above its bottom, "Jump to latest" floats over it.
Whether the thread is at its bottom is read from where the content ends
rather than from `scrollTop`, whose sign a reversed column reports
differently across engines. Content that grows after it lands (an image,
a wider window) is followed too, so the newest message never slips under
the composer.
Nothing anywhere recorded what a model call cost: the harness used a
provider's usage numbers only to decide whether a turn had finished, then
dropped them. So a reply could not say what it took to make, one model
could not be compared with another on price, and there was no way to ask
what a day, a week or a month of the crew's work came to.
- `token_usages` records one row per call: user, bot, workspace, session,
job and message, the harness, the provider and model that actually
answered, the prompt and completion counts, and the cost when the
provider reports one.
- `harness-usage-scope.ts` carries those ids in an AsyncLocalStorage
scope that `run-bot-job.ts` enters once, so a run's turns, its session
summary, its transcriptions, images and embeddings are all attributed
without threading ids through every harness function. Nested runs merge
rather than replace, and a null genuinely clears a field (a bot's own
chat has no workspace).
- Work with no run behind it is charged to its owner instead:
`run-with-fallback-scope.ts` reaches the workspace's creator else the
bot's, and deliberately does not override a real run, so an ingest a
bot asked for is not misattributed to the bot's owner.
- All six harnesses record through `withCandidates`'s `on_success`.
Recording never throws: a model that answered must not lose its answer
to a failed insert.
- A job's calls land before its message exists, so `complete-bot-job.ts`
back-fills `message_id` once the message is inserted.
- The message's actions row shows what it cost, beside the timestamp and
the play button, with the breakdown in its title.
- `/admin/usage` totals the install's spending for 7, 30 or 90 days: the
total against the previous period, a day-by-day chart, a split per kind
of call, and a table per model with the price per thousand tokens
worked out from cost and tokens, which is what makes two models
comparable. Crew managers only, since it is the whole install's work
rather than one user's conversation. It sits with Routines and Activity
in the user popover rather than under Settings, because it is a record
of what happened, not something to configure.
Days are grouped in local time (`date(created_at / 1000, 'unixepoch',
'localtime')`), matching the local midnights the window builds and the
`YYYY-MM-DD` buckets it looks up. Grouping by the bare UTC day put a late
evening's calls under a name no bucket carried, so they vanished and the
chart's total silently disagreed with the totals card; a test pins it.
`canSeeRailItem` now decides who sees a popover or rail entry. The two
each filtered with their own inline rule and only one understood
`can_view`, so a gated page could be offered in the popover to everyone.
delegate_task had no workspace argument: a task handed over from the
coordinator's own chat always ran in the worker's own chat, on that
worker's memory, files and directly assigned computers alone, even when
the work belonged to a workspace. The coordinator could not say where the
work goes, and the prompt never said the two differ.
- delegate_task takes an optional workspace_id: the worker runs the task
in that workspace's context, with its own instructions and memory
there, the project's files and the workspace's computers, and the
result lands in its conversation in that workspace.
- A worker missing from that workspace is added first (pushBotCreated)
and the reply says so.
- Handed over from inside a workspace the task always runs there; a
different workspace_id is refused instead of ignored.
- The coordinator prompt, CoordinatorReplyNote and the tool description
now spell the distinction out.
The Gitea registry is on a VPN, which a GitHub runner can never reach, so
the mirror step in the release workflow could only ever fail: its last four
releases reported a step that did nothing. The workflow now builds and
pushes to GHCR alone, which is the registry the installer defaults to, and
nothing in CI mentions Gitea.
The image reaches the Gitea registry through `scripts/mirror-gitea.sh`
(`bun run mirror:gitea`) instead, run in parallel with a release and from the
VPN side: it copies from the Gitea host over ssh, where skopeo can see both
registries, so nothing has to be installed there. `skopeo copy --all` keeps
the manifest list and its signatures, and every destination tag is read back
and compared with GHCR's digest, so a partial copy fails loudly instead of
being assumed to have worked. `latest` and `major.minor` are only moved while
that release still owns them in GHCR, so mirroring an older release never
rolls them back.
Neither direction can fail the other now, which is the point: a registry that
is down, slow or full no longer takes a release with it.
The thinking row and the header menu both carry Stop while a bot has work
open for the user anywhere: every pending or processing job of that bot in
their conversations is failed as stopped, its task closed, a Stopped step
recorded and one closing message posted per session, so the thread stops
waiting and the bot goes idle. The handler running the job is aborted from
this process through a registry, so a job stuck waiting on a model stops at
once instead of on its next heartbeat, and a late face change after the
abort is dropped.
The release stage is now said in the app too: Settings > General reports
"Cobble v0.1.5 (alpha)" next to development, and the README and RELEASING.md
say what alpha means for installs.
tbenandClaude Opus 5.5