Detection missed the npx cache, so a Playwright installed through Claude
Code's MCP setup (`npx @playwright/mcp@latest`, which unpacks into
~/.npm/_npx/<hash>/node_modules and no node_modules at all) was invisible.
The probe now globs that cache alongside the existing roots and reports
every root it consulted.
It also read `has_bind` off whichever manifest resolved first. Verified
that npm does not hoist for global installs and that the `playwright`
wrapper ships no types/types.d.ts, so `npm i -g playwright` made the pane
call a current build "predates browser.bind()". The probe now hops from
the wrapper to its nested playwright-core.
The messages no longer offer `@playwright/mcp` as a way through setup: it
bundles a playwright-core that binds but never `@playwright/cli`, so that
route could not have worked. It is named only for what it does do.
New `install.rs` + two commands do the setup, streaming on the existing
`container-progress` event and re-probing on success:
* playwright + @playwright/cli into /workspace as `claude`, --no-save.
/workspace is not a bind mount (projects mount at
/workspace/{mount_name}), so nothing of the user's is touched, no sudo
is needed, and Node resolves it from scripts in the project.
* A browser, as its own action with the size stated first: apt libraries
as root, then the download, then a real headless launch to prove it
works. The base image ships none of Chromium's shared libraries, which
is why a download could succeed and the browser still not start.
Chromium and the Chrome channel are both offered — @playwright/mcp
asks for `chrome` specifically. A certificate failure is reported as a
container trust-store problem rather than a broken install.
Installing is always user-initiated; opening the tab only probes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSP2KNPhuWKQ4DL5TZEn3k
Adversarial review of the branch produced findings across four areas.
This addresses them, plus the Windows CI environment.
Secrets. commit_container_snapshot baked the container's full env into
the per-project snapshot image, so the shared OAuth token — and the AWS
keys, git token and gateway master key — outlived revocation and were
readable via docker inspect. Verified against Engine 29.6 that a commit
body's config merges over the container's: keys cannot be dropped but
can be overwritten, so all of them now commit as KEY=. clear_claude_token
additionally rewrites images from earlier builds and reports honestly
when a tag could not be rewritten.
The recommendation to move the token out of env entirely was not taken,
with reasoning: apiKeyHelper is a different auth method that outranks
CLAUDE_CODE_OAUTH_TOKEN rather than a transport for it, and no
file-based delivery exists. The durable exposure — the image — is what
is closed here. Separately noted, not fixed: entrypoint.sh captures the
token into the scheduler's .env inside the persisted volume.
URL spoofing. Three call sites reached openUrl with container-controlled
strings, one of which the review missed (the WebLinksAddon handler).
The sign-in URL was scraped from container output with a longest-match
tie-break and no userinfo check, so claude.ai@evil.tld rendered as
"claude.ai…" in a truncating element. There is now one sanitizer in
front of every sink — scheme allowlist, no userinfo, C0/C1 and quote
rejection, host allowlist for the sign-in case, first-match — and the
origin renders un-truncated. The toast is keyed so a changed URL
remounts, closing a bait-and-switch where the user read one URL and
clicked another.
Migration. The rollback pin was best-effort: a tag failure was logged
and the migration continued past remove_container, after which the
final commit overwrote the only copy of the old system layer. It now
aborts before anything destructive and reads the tag back. /var was
destroyed while the ordinary recreate path preserves it — making the
"safe" alternative to Reset more destructive than Reset's alternative;
data-bearing subtrees are now detected and disclosed in the pre-flight
rather than copied, since tarring a live database onto a different
base's packages is a corruption risk. resume_migration now verifies the
migration-state label instead of reporting success for a container that
never swapped. dismiss actually resolves the record rather than leaving
the feature permanently refusing to migrate. Start and Reset are guarded
while a migration is live.
Lifecycle. The gateway no longer publishes on 0.0.0.0 — bind address and
advertised URL are derived together so they cannot drift. Disabling it
now stops it. App exit runs teardown concurrently under a budget with a
visible shutting-down state instead of blocking for minutes. Auto-starts
retry when Docker is not up yet, and the polling-recovery path now
reconciles, so interrupted migrations are still recovered. Auth-bridge
forwards are capped, closing a container-driven fd exhaustion.
Windows CI. build-windows failed on this branch with "linker link.exe
not found". The runner had no MSVC build tools and the workflow assumed
a hand-provisioned machine, so a bare runner registers, accepts jobs and
fails at link time after downloading the whole crate graph. The job now
installs the VC++ workload when vswhere cannot find it, matching how it
already conditionally installs Rust and Node.
192 Rust tests, 274 frontend tests, both builds clean, zero warnings.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Projects were pinned to the image they were first created from. Both
create paths preferred triple-c-snapshot-<id>:latest whenever it
existed, and container_needs_recreation compared the container's live
image against the triple-c.image label — which create_container wrote
from the same image it created from. A tautology that could never fire.
The only escape was Reset, which calls remove_project_volumes and
destroys the login, skills and transcripts.
Measured consequences on this host: real projects are missing socat (so
the auth bridge cannot tunnel) and bubblewrap (so sandbox mode does not
work), plus Mission Control and triple-c-sso-refresh, and sit 61
packages behind the base including ca-certificates, openssl and curl.
Detection. create_container now writes triple-c.base-image-id (the image
ID, not RepoDigests, which local-built and custom images do not have)
and triple-c.create-image. container_needs_recreation takes the expected
create-image and compares against the latter, so the check means
something. base-image-id is deliberately NOT compared: a base bump would
otherwise silently recreate from the snapshot, consuming the "you should
migrate" signal without migrating. Staleness is surfaced, never acted on
automatically.
Migration keeps the volumes. /home/claude and ~/.claude are volumes and
the image's copy is seed-only — permanently masked after first mount —
so the login, ~/.claude.json, skills, transcripts, scheduler tasks, SSH
keys, cargo, uv, ruff and Claude Code itself re-attach untouched. Only
root-level state is rebuilt: apt packages are replayed against the new
base rather than copied, so no stale libc is dragged forward, and
/usr/local, /opt and the non-bind-mounted parts of /workspace are copied
verbatim with tar --skip-old-files so they can never clobber a newer
base binary.
docker diff is not used: on a snapshot-derived container it reports only
changes since the last commit. Raw image-vs-image diffing is filtered
through dpkg ownership because it otherwise lies — 8,677 raw path
differences on a real project reduced to 2 genuinely user-authored
files, both loose /workspace-root files.
Crash safety. snapshot:latest keeps pointing at the old image until the
final commit, so any crash before it self-heals on next start. Later
crashes are caught by reconcile_project_statuses. The rollback pin is a
docker tag: 0.057s and 0 bytes. Rollback restores the system layer only
— volumes are never touched — and the UI says so rather than implying a
time machine.
Fixes an infinite recreation loop shipped with the MCP removal. docker
commit propagates labels to the image, so a container created from a
snapshot inherited its non-empty triple-c.mcp-fingerprint and the
one-shot shim recreated it again on every start, forever. Lineage labels
are now always written explicitly.
Documents the second, separate bug this uncovered: Dockerfile changes
under /home/claude never reach an existing project, migration or not,
because the volume masks them. Anything that must stay upgradable
belongs in /usr/local/bin or /opt, or must be seeded by entrypoint.sh.
145 Rust tests, 227 frontend tests, both builds clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Four features, plus a latent bug fix.
llama.cpp backend. Claude Code only ever speaks the Anthropic Messages
API — confirmed empirically by pointing it at a logging server, which
received POST /v1/messages?beta=true. llama-server implements that
natively (verified in its README, alongside --port default 8080), so
this is a plain base-URL backend with no translation shim, the same
shape as Ollama. Its --api-key defaults to none, so the auth token is a
placeholder Claude Code requires and llama-server ignores.
Model alias fix. ANTHROPIC_DEFAULT_HAIKU_MODEL is documented as "also
used for background functionality", and Triple-C set none of the alias
vars. So on every custom-endpoint backend, Claude Code resolved `haiku`
to an Anthropic model id and sent it to a local server that does not
have it — background features failed silently. All four
ANTHROPIC_DEFAULT_{OPUS,SONNET,HAIKU,FABLE}_MODEL vars are now pinned to
the backend's configured model, with an optional Haiku override, and
blanked for Anthropic and Bedrock so those keep Claude Code's defaults.
The deprecated ANTHROPIC_SMALL_FAST_MODEL is never emitted. Existing
Ollama and OpenAI-Compatible containers are recreated once so the new
env reaches them; the snapshot is preserved.
Model gateway. Optional LiteLLM sibling container, off by default,
mirroring stt.rs — this is what makes real OpenAI usable, since
api.openai.com has no /v1/messages. Pinned to v1.96.0 by tag and digest:
the 1.82.7/1.82.8 malware was PyPI-only and never affected the official
images, which is precisely why this builds FROM the image rather than
pip-installing, but 1.84.0 is still the floor for proxy CVEs (API-key
SQLi, Host-header auth bypass, MCP auth bypass). Binds 0.0.0.0 because
project containers consume it, and therefore always sets a master_key —
LiteLLM without one accepts any key. The provider key lives in the OS
keychain and is uploaded into a volume, never an image layer or label.
URL relay. A container-side xdg-open/BROWSER shim opens URLs in the
host's browser. Uses an OSC sequence to /dev/tty rather than a printed
sentinel, because the shim usually runs as a grandchild of a process
capturing its children's output. Degrades to printing the URL when no
terminal is attached, so scheduled tasks do not hang. Only http/https,
with control characters rejected before new URL() — which strips
newlines, so java\nscript: would otherwise parse as javascript:. Nothing
auto-opens; the user confirms. The web terminal shows a tap-to-open
banner instead, since that browser may be a phone across a tunnel.
Browser view. A Project Home tab that watches and takes over the browser
Claude drives with Playwright, using Playwright's own dashboard. Zero
image cost — Playwright stays user-installed. It does not reuse the auth
bridge's PortForward, which binds an unauthenticated port: correct for a
throwaway OAuth listener, wrong for mouse and keyboard control of a
browser in a passwordless-sudo container. Instead a token-gated loopback
proxy checks Host, then token or a forbidden-header origin signal,
before a byte reaches the container. Host ports are confined to
47820..=47827 so CSP frame-src can enumerate them rather than widening
to a wildcard, with a test asserting the two agree.
188 frontend tests, 107 Rust tests, both builds clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Completes the Automation tab: it could list, toggle, run, log and remove
tasks but not create them, so task creation still meant dropping to the
CLI. Adds add_scheduled_task and update_scheduled_task, plus a task
editor with cron presets and a plain-English reading of the expression.
Every field is free user text, so all of it goes to the scheduler as a
bare argv vector through bollard — no shell, no quoting. Validation is
shape-only rather than metacharacter scrubbing: length caps, no control
characters in single-line fields, no leading-dash name, absolute
working_dir. Verified by round-tripping a prompt containing
`; rm -rf /`, `$(id)`, backticks and newlines: it landed byte-for-byte
in the task JSON with nothing executed.
The scheduler CLI has no `edit`, so update is add-then-remove with the
add first — a rejected edit leaves the original intact. The new id is
surfaced in the editor rather than hidden.
Root-cause fix, and the more serious half of this commit:
triple-c-scheduler never validated --schedule, and rebuild_crontab
regenerates the entire crontab and pipes it to `crontab`, which rejects
the whole file if any line is malformed — with the error thrown away by
`2>/dev/null || true`. A single bad schedule therefore silently
unscheduled every other task in the container while reporting success.
Reproduced directly. It matters because the global CLAUDE.md tells
Claude to drive this CLI, so Claude could trigger it unprompted.
`add` now validates the expression and exits non-zero, and
rebuild_crontab reports a rejected crontab instead of swallowing it,
keeping the offending file for inspection. Verified against the real CLI
in this container: a bad schedule is refused without disturbing an
existing task's crontab entry, and `0 9 * * 1-5`, `*/30 * * * *`,
`0,30 8-17 * * *` and `0 0 1 1 *` are all still accepted. The Rust layer
validates independently, agreeing with vixie cron on 23 probed
expressions including `1/2` and `*/0` being invalid.
121 frontend tests, 44 Rust tests, both builds clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Reset is destructive in a way its name does not advertise:
rebuild_project_container deletes both project volumes, so it wipes the
claude login, anything installed in the container, and every saved
session transcript. It was a single unconfirmed click in the overflow
menu, while the comparably destructive Remove already confirmed. Adds
ConfirmResetModal, which names each loss and says explicitly that the
host-side mounted folders are untouched.
Docs: the user guides still described the pre-Project-Home UI. Sixteen
factually wrong statements corrected, including "expand the Config
panel" (six sites), the actions table (Reset and Remove are in an
overflow menu, Files is a tab), a progress modal that no longer exists,
a double-click-to-rename gesture ProjectRow never had, the Full
Permissions boolean, an incomplete reserved-env list, and the claim in
TECHNICAL.md that OAuth tokens survive a Reset. Both layout diagrams and
the project tree were rebuilt from the filesystem.
New sections cover permission modes with the exact CLI mapping, Project
Home, Sessions, capability tiles, Automation, shared authentication, the
Auth Bridge and its security posture, and keyboard shortcuts.
Known gap recorded rather than papered over: the Automation tab manages
existing scheduled tasks but cannot create them — no add command is
registered — so task creation remains `triple-c-scheduler add` in the
terminal.
87 frontend tests, 34 Rust tests, both builds clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
UI for the shared Claude token: a Settings section showing token state
with Authenticate and Revoke, an acquisition modal built on the shared
Modal (sign-in link handed to the host browser via the opener plugin,
plus the code input that answers `setup-token`'s stdin prompt — the flow
cannot complete without it), and a per-project opt-out toggle shown only
for the Anthropic backend.
Cancellation: acquire_claude_token previously had only two exits,
completion and a 15-minute timeout, and held the single-flight guard for
the whole time. Closing the dialog therefore locked the user out of
retrying for up to 15 minutes. Adds cancel_claude_token, backed by a
oneshot claimed and released in lockstep with the input guard, selected
on in the run loop so it wins the race and tears the exec down. The
dialog's Cancel now calls it and closes either way.
Also refreshes CLAUDE.md, which had drifted: it documented the deleted
ProjectCard, and asserted that new IPC commands need permission grants
in capabilities/default.json — they do not, that file covers plugin
commands only. Adds the conventions that would otherwise bite:
container_needs_recreation() is purely label-based and never diffs env,
so container-affecting state needs its own label; and #[serde(default)]
on a bool yields false regardless of intent.
Corrects the claim that Reset preserves credentials. Reset calls
remove_project_volumes, which deletes both the home and claude-config
volumes, so it wipes ~/.claude, the OAuth token, installed skills and
session transcripts.
84 frontend tests, 34 Rust tests, both builds clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Project Home (DESIGN-REVIEW §B2): the project is promoted from a 280px
sidebar card to a first-class main-area view. ProjectCard.tsx (1,257
lines) is replaced by a select-only ProjectRow plus tabs for Overview,
Sessions, Automation, Config and Files. The PortMappings, FileManager
and ContainerProgress modals are absorbed rather than reimplemented.
Config gains a Saved/Saving/Failed indicator — save-on-blur failures
previously reached only console.error.
Tier-1 polish (DESIGN-REVIEW §A): new elevation, muted-accent, disabled
and focus-ring tokens; a global :focus-visible ring with every
focus:outline-none removed; filled buttons moved to --accent-emphasis
and white-on-success toggles retired, fixing three WCAG AA failures
(2.1:1, 2.5:1, 2.4:1); a shared Modal primitive with role="dialog",
focus trap and restore, adopted by all remaining modals; status
indicators that carry a glyph and word rather than colour alone.
Ctrl+Shift+W closes a tab, deliberately not Ctrl+W — that is readline's
kill-word, used constantly in the terminal this app is built around.
Auth Bridge: a general loopback-callback bridge so browser logins run
inside a container (aws sso login, Concourse fly login, claude login)
can complete against the host browser. Listeners are discovered from
/proc/net/tcp{,6} — ss/netstat/lsof are absent from the image — bound on
host 127.0.0.1 only, and tunnelled in over the Docker API via socat,
which keeps working on Docker Desktop where container IPs are not
routable. Falls back to [::1] because Node resolves localhost to IPv6
first, so claude login often binds ::1 alone. Opt-in per project.
This extracts create_attached_exec() and moves the existing terminal
session path onto it, so there is one attached-exec implementation
rather than two.
Shared auth token: `claude setup-token` is run in a container, the token
is stored in the OS keychain and injected as CLAUDE_CODE_OAUTH_TOKEN
into Anthropic-backend projects. Contrary to the initial design note,
setup-token uses an Anthropic-hosted redirect and blocks on a stdin
paste prompt rather than a loopback callback, so a stdin command is
required for the flow to complete.
The token is never logged, never returned to the frontend, and is
redacted from the streamed output with a stateful matcher that withholds
any tail that could still grow into a secret. Change detection uses a
random rotation id rather than a hash, since a hash in a docker-inspect
readable label would be an offline verification oracle.
Frontend 33 -> 51 tests; Rust 34 tests. Both builds clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Permission modes: replaces the binary full_permissions flag with a
PermissionMode enum (Plan/Default/AcceptEdits/Bypass). Flag mapping is
defined once in PermissionMode::cli_args() and used by the terminal, the
web terminal, and the scheduler:
Plan -> --permission-mode plan
Default -> (no flag)
AcceptEdits -> --permission-mode acceptEdits
Bypass -> --dangerously-skip-permissions
Choices verified against `claude --permission-mode` on 2.1.226.
full_permissions is retained and effective_permission_mode() falls back
to it, so existing projects.json needs no migration.
Bug fix: triple-c-task-runner ran `claude -p ... --dangerously-skip-
permissions` unconditionally, ignoring the project's setting entirely.
It now reads TRIPLE_C_PERMISSION_MODE, which is injected into the
container, added to the reserved env blocklist, propagated through the
entrypoint's cron env filter, and tracked by a new
triple-c.permission-mode label so a change forces recreation.
Introspection: new commands/inspect_commands.rs exposes read-only views
into the container over docker exec — Claude sessions (parsed from
~/.claude/projects/<cwd>/<uuid>.jsonl), installed capabilities (skills,
agents, commands, hooks, plugins, natively-configured MCP servers), and
the triple-c-scheduler task list, logs and notifications.
Task/session ids are validated against a strict allowlist and every
parameterized call runs as a bare argv vector via bollard, so no shell
is involved. Stopped containers return empty results rather than errors.
No UI yet; that lands with the Project Home view.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude Code manages MCP natively now (`claude mcp add/list/remove`,
`.mcp.json`, `/mcp`), so Triple-C's own MCP server library is redundant.
Deletes components/mcp/, hooks/useMcpServers.ts, the MCP sidebar tab and
rail icon, the per-project enable checkboxes on ProjectCard, the
mcpServers slice of the Zustand store, the four IPC wrappers, and the
McpServer/McpTransportType types.
Rust backend is untouched in this commit; the commands simply become
unreachable. Backend removal and the legacy container/network cleanup
follow separately.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
- F2: upload_host_file_to_container now reads the dropped file into a Vec
inside the blocking task and sizes the tar entry from those exact bytes,
rather than stat-then-stream where a file changing size between the
stat and the read could desync the tar header and silently corrupt the
archive. Still runs off the async worker; memory stays bounded by the
256 MiB drop cap.
- F4: the "Backup saved" confirmation now auto-clears after 8s (guarded
against clobbering a newer status message) instead of lingering in the
project card's status line indefinitely.
F1 (claimed AWS CLI regression from empty-env neutralization) was a false
positive: verified against aws-cli 2.35 that an empty AWS_ACCESS_KEY_ID is
treated as absent and botocore falls through to ~/.aws/credentials (the
call reached AWS and returned InvalidClientTokenId for the file's key, not
PartialCredentialsError). No change needed.
cargo check / tsc / vitest all pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- L-e: route terminal file drops purely by a bounds hit-test instead of
the `active` flag. Inactive panes are display:none (zero-size rect) so
they never match; a zero-size guard makes that explicit. Correct for
the current tabbed layout and future-proof for split panes, where a
drop on a visible-but-unfocused pane previously matched no handler.
- L-f: stream the dropped file straight into the upload tar inside a
blocking task (new exec::upload_host_file_to_container) instead of
reading the whole file into a Vec and then re-packing it. Peak memory
drops from ~2x to ~1x the file size, and the synchronous file IO no
longer runs on the async worker.
cargo check / tsc / vitest all pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- L4: sync_bedrock_credentials (renamed from write_bedrock_static_
credentials) now also clears a stale ~/.aws/credentials when the
project no longer uses static-credential Bedrock, so static keys don't
linger unused in the persistent home volume after switching backends.
Skipped when /tmp/.host-aws is mounted (host-managed ~/.aws). HOME is
also set explicitly on the exec env for robustness.
- M1: the Backup button now has a tooltip and the success toast notes
that the archive includes MCP/config which may contain MCP-embedded
API keys (OAuth tokens are excluded) — keep it private.
- L2: backup now uses async file IO (tokio::fs::File + AsyncWriteExt,
tokio::fs::remove_file) instead of blocking std::fs between awaits;
dropped-file reads use tokio::fs::metadata/read.
- L5: upload_host_file_to_terminal explicitly `mkdir -p`s
/tmp/triple-c-drops instead of relying on Docker's tar extractor to
create the parent dir.
Verified L4 cleanup guard, L5 mkdir, async IO, and exit-code paths
against real containers. cargo check / tsc / vitest all pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fixes from the code review of this branch:
- Backup requires a running container (it runs via `docker exec`, which
can't run on a stopped one). Removed the misleading "Backup" button
from the stopped-project actions, added an explicit running check with
a clear error, and corrected the doc comment. (H1)
- jq sanitization fallback no longer leaks secrets: if ~/.claude.json
can't be parsed, the backup substitutes an empty object and warns to
stderr instead of copying the raw file (which held primaryApiKey /
oauthAccount). Verified the raw key never reaches the archive. (H2)
- Dropped-file paths typed into the terminal are now always single-quoted
(with '\'' escaping), not only when they contain whitespace — a name
like `foo$(whoami).txt` was previously sent raw into the shell. (M2)
- write_bedrock_static_credentials checks the exec exit code via the new
exec_oneshot_env_status and fails loudly on a write/chmod error instead
of silently reporting success. exec_oneshot keeps its
ignore-exit-code behavior so list_container_files is unaffected. (M4)
- Backup removes a partial/truncated archive on any stream error and
treats a non-zero tar exit code as failure (a truncated gzip was
previously reported as success). (L1)
- Dropped files are capped at 256 MiB to avoid ballooning host RAM
(the file is read fully into memory then re-tarred). (M3)
- Stopped excluding .git/objects from the backup so git history,
including unpushed commits, is preserved faithfully. (L3)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Drop files onto a terminal pane and they're copied into the container and
their in-container paths typed into the prompt, so Claude Code can read
them for reference — mirroring the existing image-paste flow.
Backend: upload_host_file_to_terminal reads the dropped host file and
writes it under /tmp/triple-c-drops/<name> in the session's container,
returning that path. Rejects directories and unreadable paths.
Frontend: TerminalView subscribes to Tauri's webview onDragDropEvent
(OS file drops are intercepted at the webview level, so HTML5 ondrop
wouldn't expose paths). The window-wide event is guarded by the pane's
`active` flag plus a bounds hit-test so a drop only affects the terminal
it landed on; multiple files are uploaded and their paths inserted
space-separated (quoted when they contain spaces).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds a manual backup button on each project card (next to Start/Reset
when stopped, and next to Files when running) that saves a gzipped
tarball of the container's /workspace to a host path via the native save
dialog.
Backend: download_container_backup runs `tar czf -` inside the container
(so excludes + compression happen there rather than streaming a 16 GB
workspace) and pipes stdout straight to the chosen file. Regenerable
build artifacts (node_modules, target, .git/objects) are excluded so the
archive stays restore-sized. Returns bytes written; stderr is captured
for error reporting and a zero-byte result is treated as failure.
Works whether the container is running or stopped (only requires that it
exists). Verified on the Ubuntu/GNU-tar container base.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-up to PR review on terminal-layout-statusbar:
- [Major] Pin STT transcripts to the originating terminal. The single
useSTT instance is bound to the live active session, which can change
mid-recording. Capture the session id at recording start in a ref and
inject the transcript there instead of the live sessionId, so text
always lands in the terminal where recording began.
- [Minor] Clear the status-bar scroll state when the active terminal
unmounts, and null out termRef on dispose, so scrollActiveToBottom
can't point at a disposed terminal. Tab switches don't unmount, so
this only fires when the active session is actually closed.
- [Nit] Fix the terminal padding comment to match the symmetric value.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Terminal layout fixes for the xterm pane:
- Stop the terminal grid from clipping its rightmost column / bottom
row. The padding was on the element xterm mounts into, which the
FitAddon measures; the grid overhang got clipped. Padding now lives on
a wrapper and the xterm host fills it with no padding.
- Move the STT mic from a floating bottom-left overlay into the status
bar (far right). A single useSTT instance bound to the active session
now lives in App; Ctrl+Shift+M routes through the store.
- Move "Jump to Current" from a floating terminal overlay into the
status bar. The active TerminalView surfaces its scroll state and
scroll action via the store.
- Tighten terminal padding (was 8/12/48/16) now that nothing floats over
it, so the terminal claims as much area as possible.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a time-bounded `claude update` to entrypoint.sh that runs as the
claude user before the container is marked ready, so every terminal
session launches the latest CLI. Non-fatal and capped at 120s so an
offline/slow network never blocks container readiness; PATH covers both
~/.claude/bin and ~/.local/bin install locations.
Add flex-shrink-0 to the FileManagerModal header/footer so a long file
list can't squeeze them and the scroll region stays robust.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Right-click a tab (or double-click) to rename. Renamed labels show
as "ProjectName: CustomName" and are stored in the project's
renamed_session_names map. The entry is cleared on tab close.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New fields: GlobalAwsSettings.default_model_id, plus
GlobalOllamaSettings and GlobalOpenAiCompatibleSettings (base_url +
default_model_id each). When a per-project base_url or model_id is
blank, the container env vars and config fingerprints fall back to
the global value. Container recreation is triggered whenever the
resolved value changes, so editing a global default updates existing
projects on next start.
UI: added the new fields to AwsSettings and two new global settings
components, slotted into the Backends accordion.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Multiple-open accordion with per-section state persisted to
localStorage. Sections: General, Backends, Container, Git/SSH,
Tools, Updates. General is open by default; the rest are collapsed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Persist collapsed state in localStorage. When collapsed, render a
narrow rail with Projects/MCP/Settings icon buttons that expand the
sidebar to that view on click.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sandbox mode: new per-project toggle that turns on Claude Code's bash
sandbox inside the container. Adds `bubblewrap` and `socat` to the
Dockerfile (the two Linux deps required by the sandbox), and emits a
managed `sandbox` block into `~/.claude/settings.json` via the existing
CLAUDE_CODE_SETTINGS_JSON entrypoint merge:
- `enabled` mirrors the Triple-C toggle and is always emitted, so the
entrypoint's recursive jq merge clears any prior on-state from the
persisted named volume — Triple-C is authoritative.
- `enableWeakerNestedSandbox: true` because we run inside Docker without
privileged user namespaces.
- `allowUnsandboxedCommands: false` to disable the `dangerouslyDisableSandbox`
escape hatch — opting into the sandbox shouldn't come with a runtime
bypass.
When sandbox is on, a SANDBOX_INSTRUCTIONS section is appended to
CLAUDE_INSTRUCTIONS so Claude can guide users through allowing extra
paths/domains, excluding `docker *`/`watchman *` from the sandbox, and
the rule that `sandbox.enabled` is owned by Triple-C. The Claude-Code
settings fingerprint includes sandbox state (only when on, to avoid
spuriously flagging existing containers for recreation on upgrade).
Bedrock service tier: new optional field on the per-project Bedrock
config. When set, exported as ANTHROPIC_BEDROCK_SERVICE_TIER (added in
Claude Code 2.1.122) and included in the Bedrock fingerprint.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When Docker isn't detected on startup, surface a dialog offering a
one-click install (pkexec + get.docker.com on Linux, brew cask on
macOS, winget on Windows) with a graceful fallback to manual steps
and a link to official documentation. Install output streams back
to the UI via a tauri event.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add bottom padding to terminal containers so FitAddon proposes one
fewer row, leaving visible space below Claude Code's mode indicator.
Previously the bottom status line (e.g. "bypass permissions on") was
clipped against the container edge in fullscreen TUI mode.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds first-class support for Claude Code CLI features (2.1.71-2.1.110):
- New ClaudeCodeSettings struct with per-project and global defaults for
TUI mode, effort level, focus mode, thinking summaries, session recap,
auto-scroll, env scrub, and 1-hour prompt caching
- Settings injected as env vars (CLAUDE_CODE_NO_FLICKER, etc.) and
~/.claude/settings.json entries via entrypoint.sh merge block
- New ClaudeCodeSettingsModal component for configuring settings
- Session naming support (-n flag passed to claude CLI, shown in tabs)
- Relaxed reserved prefix filter: CLAUDE_CODE_* env vars now allowed in
custom env vars UI for power users
- Global SSH key path, git name, and git email now used as fallbacks
when per-project values are not set, with UI in SettingsPanel
- Fingerprint-based change detection triggers container recreation when
Claude Code settings change
- Updated README, HOW-TO-USE, and CLAUDE.md documentation
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replaces the native title attribute with a custom tooltip that appears
instantly on hover, displaying the shortcut in a styled kbd element.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Lifts useSTT hook from SttButton into TerminalView so both the hotkey
and the button share the same recording state. The hotkey keeps terminal
focus so after transcription the user just presses Enter. The button
also no longer steals focus via onMouseDown preventDefault.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds a mic button to the terminal UI that captures speech, transcribes
it via a Faster Whisper sidecar container, and injects the text into
the terminal input. Includes settings panel for model selection
(tiny/small/medium), port config, and container lifecycle management.
- stt-container/: Dockerfile + FastAPI server for Whisper transcription
- Rust backend: STT container management, transcribe_audio IPC command
- Frontend: useSTT hook, SttButton, SttSettings, WAV encoder
- CI: Gitea Actions workflow for multi-arch STT image builds
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds an axum HTTP+WebSocket server that runs alongside the Tauri app,
serving a standalone xterm.js-based terminal UI accessible from any
browser on the local network. Shares the existing ExecSessionManager
via Arc-wrapped stores, with token-based authentication and automatic
session cleanup on disconnect.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The onScroll RAF optimization (only fire when atBottom changes) prevented
the button from showing because xterm's onScroll may not fire from wheel
events. Fix by setting isAtBottom(false) directly in the wheel handler
and removing the RAF guard to always schedule state updates.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Prevent viewport jumping during Claude output by only re-enabling
auto-follow on user-initiated scrolls (wheel events within 300ms),
not on write-triggered xterm scroll events. Add a "Following/Paused"
toggle button in the top-right corner of the terminal.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The previous fix checked isAtBottomRef inside the write callback, but
xterm's own scroll events during write processing could set the ref to
false (viewport desync), breaking auto-follow entirely.
Introduce a separate autoFollowRef that tracks user intent:
- Set to false only by explicit mouse wheel scroll-up (capture phase)
- Set to true when viewport reaches bottom or user clicks the button
- Write callback uses autoFollowRef so desync doesn't kill auto-follow
but user scroll-up correctly pauses it
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The shouldFollow flag was captured before term.write() but the callback
ran asynchronously — if the user scrolled up in between, the stale flag
forced the viewport back to bottom, preventing the button from appearing.
Check isAtBottomRef at callback time instead so user scroll-up is respected.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Auto-scroll viewport on new output when user is at bottom, debounce
scroll state updates to reduce re-renders, preserve scroll position
across resize reflows, and fix "Jump to Current" button by re-fitting
the terminal to clear viewport desync.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- help_commands.rs: fetch HOW-TO-USE.md from GitHub raw instead of Gitea
- DockerSettings.tsx: display GHCR image address in settings UI
- HOW-TO-USE.md: update registry description to ghcr.io
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
New projects default to standard permission mode (Claude asks before acting).
Existing projects default to full permissions ON, preserving current behavior.
UI toggle uses red/caution styling to highlight the security implications.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reflects that this backend works with any OpenAI API-compatible endpoint
(LiteLLM, OpenRouter, vLLM, text-generation-inference, LocalAI, etc.),
not just LiteLLM. Includes serde aliases for backward compatibility with
existing projects.json files.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When the user highlights text in the terminal, a "Ctrl+Shift+C to copy"
hint appears in the status bar next to the project/terminal counts.
The hint disappears when the selection is cleared.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Ctrl+C in the terminal sends SIGINT which cancels running Claude work.
This adds a custom key handler so Ctrl+Shift+C copies selected text to
the clipboard without interrupting the container.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The model field must be set and the model must be pre-pulled in Ollama
before the container will work. Updated README, HOW-TO-USE, and the
ProjectCard UI label/tooltip to reflect this.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Help dialog now fetches HOW-TO-USE.md live from the gitea repo on open,
falling back to the compile-time embedded copy when offline. Content is
cached for the session. Removes the ~600-line hardcoded markdown constant
from HelpDialog.tsx in favor of a single source of truth.
Adds a Table of Contents with anchor links for quick navigation and a new
troubleshooting entry for the "Failed to install Anthropic marketplace"
error with the jq fix. Markdown renderer updated to support anchor links
and header id attributes.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Rewrite Tooltip to use React portal (createPortal to document.body) so
tooltips render above all UI elements regardless of ancestor overflow:hidden.
Also increased max-width from 220px to 280px for longer descriptions.
Expanded Backend tooltip to explain each option (Anthropic, Bedrock,
Ollama, LiteLLM) with practical context for new users.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add circled ? help button in TopBar that opens a dialog with HOW-TO-USE.md content
- Create reusable Tooltip component with viewport-aware positioning
- Add 32 tooltip indicators across project config and settings panels
- Covers backend selection, Bedrock/Ollama/LiteLLM fields, Docker, AWS, MCP, and more
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Fix serde deserialization error: TypeScript sent "lit_llm" but Rust expected "lite_llm"
- Rename AuthMode enum to Backend across Rust and TypeScript (with serde alias for backward compat)
- Add container image update checking via registry digest comparison
- Improve Settings page: fix image address display spacing, remove per-project auth section
- Update UI labels from "Auth" to "Backend" throughout
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
