My previous fix made submit no longer block on isStarting, with the
goal of keeping local quick-chat sessions from silently dropping
messages. The trace from CI showed it traded one failure for another:
the editor became contenteditable while the session was still
STARTING, the e2e suite pressed Cmd+Enter immediately, and the
backend rejected with "Failed to send message to agent" because the
agent process wasn't ready yet.
Restore the original semantics — STARTING blocks both editor and
submit — but keep `isPreparingEnvironment` as the narrower trigger for
the "agent still being set up" tooltip. Net behaviour:
- STARTING transition: editor disabled, no tooltip (signal is the
disabled state itself)
- container/sandbox prepare: editor disabled + tooltip explaining why
- quick-chat sessions: editor stays disabled until RUNNING, no race
The "agent still being set up" submit gate was triggering on every
session that briefly transitions through STARTING — including local
quick-chat, which has no prepare phase and has its agent live within
~100ms. The eager-init quick-chat e2e tests (chat/quick-chat.spec.ts
and chat/message-queue.spec.ts) wait for the editor to become
editable, then press Cmd+Enter. The race made wrappedSubmit a no-op
exactly when the test was firing — message dropped, agent never saw
it, downstream assertions failed.
Split the merged isStarting flag in useSessionState: keep the broad
flag (still used for the workspace-prep loading UI), and expose
isPreparingEnvironment as a separate signal that's only true while
prepareProgress.status === "preparing". Thread it through chat-input-
area → chat-input-container → useChatInputContainer and gate the
submit-disabled tooltip on it instead of the broad isStarting.
Net: Docker / Sprites bootstrap still surfaces the "agent still being
set up" message and disables submit; quick-chat / local sessions stop
silently dropping in-flight messages.
CI's only-new-issues lint surfaced three real items left over from
recent additions:
- container.go bootstrap snippet had two `fi` shell tokens flagged by
dupword. Tagged with the same nolint directive used by the
KandevBranchCheckoutPostlude (two distinct shell blocks closing).
- DockerExecutor.CreateInstance was 89 lines (over funlen 80). Split
the body into helpers: reportCreateInstanceProgress (deferred
progress reporter), tryReconnect (gates the resume path),
seedSessionDir (per-container auth seed), buildContainerLaunchConfig,
and buildCreatedInstance. CreateInstance is now a 30-line top-down
orchestrator.
- executor_execute.go was 887 effective lines (over revive 800). Move
reuseExistingEnvironment + its helpers (latestExecutorRunningForEnv,
executorRunningMatchesEnvironment, applyExecutorRunningMetadata,
ensureLaunchMetadata) into a sibling executor_environment_reuse.go.
Resolves conflicts from concurrent main work:
- helpers.go: keep our 4-arg RegisterDockerRoutes (with task-title
provider) and bring in main's Slack route registration block.
- service_turns_test.go: keep both lifecycle and events imports.
- kanban-card.tsx: take main's slim version — main extracted
KanbanCardContextMenu / KanbanCardShell / kanban-card-content into
sibling files which the new tree depends on. Our PR's executor
visualization (RemoteCloudTooltip) lives in kanban-card-content.
- passthrough-terminal.tsx + use-passthrough-terminal.ts: keep our
onDisconnected callback alongside main's manualInputRouting and
onWsReady (mobile keyboard input routing). Adopted main's
teardownWebSocket helper and dropped the now-unused
cleanupTerminalConnection.
- event_handlers_linear.go: switch to existing
models.MetaKey{Agent,Executor}ProfileID constants so the merged
diff stays under the goconst threshold.
* fix(web): rank longest/quickest tasks by active duration
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* refactor(web): drop redundant sorted alias in TaskDurationList
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---------
Co-authored-by: Kandev Agent <agent@kandev.dev>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
The recovery branch (Failed execution → CleanupStaleExecutionBySessionID
→ relaunch) is the load-bearing piece of shell session resilience and
had no direct test. Extract the body into a free function that takes a
narrow shellLifecycle interface so the four branches are testable
without standing up a real lifecycle.Manager:
- happy path (running execution, no cleanup)
- stale recovery (Failed → cleanup → fresh)
- initial GetOrEnsureExecution failure
- cleanup failure must NOT trigger a re-launch loop
- relaunch failure surfaces a clear "recover execution" error
Address recurring review feedback on the docker reconnect path:
- reconnectToContainer was 98 lines (over the 80-line repo limit). Split
into ensureContainerRunning (inspect + start + re-inspect) and
bringupAgentctl (health-check, find-or-create instance, transparent
re-handshake). The orchestration body now reads top-down at 49 lines.
- resolveDockerEndpoint and waitForAgentctlHealth grew real branches in
this PR (host-port lookup with container-IP fallback; 240×500ms
retry loop) but had no direct tests. Introduce narrow hostPortLookup
/ healthChecker interfaces, satisfied by *docker.Client and
*agentctl.ControlClient, so both functions are now driven by stubs
that exercise happy-path, fallback, retry-until-healthy, exhaustion,
and context-cancellation branches.
- Add direct unit tests for the docker package's pure helpers
(normalizeDockerHostIP, buildDockerPortBindings) and for
dockerAgentctlPortBindings — a regression in the published port
range or host IP would silently break container reconnect.
CodeQL flagged go/path-injection on the localFileUploader.WriteFile
sink: instanceID flows into the write path, and a malformed value
would let writes escape the kandev session root and clobber arbitrary
host files. Bind the uploader to its session root and reject any path
that filepath.Rel resolves to a "../" escape. Production callers
always pass an internally generated instanceID; the check is defense
in depth against a future caller plumbing in user input.
Move SESSION_ID and friends above the vi.mock factories so the mock
closures don't depend on vitest hoisting magic capturing TDZ-undefined
locals. The previous layout worked because factories aren't invoked
until the module under test is imported (after const initialization),
but a refactor that imports earlier would break the closure silently.
Replace `plan.index(spriteStepWaitHealthy) == 1` with an
`assertPlanOrder` helper that checks "X runs before Y". The contract
the test guards is the order of create-sprite → wait-healthy →
agent-instance, not their absolute slot numbers — adding a new step
shouldn't break the test if the order is preserved.
Each piece of the resume-fast path is unit-tested separately
(StopInstance keeps the token cache, spritesShouldReconnect reads
sprite_name, the reconnect plan has 3 steps). This new test wires them
together end to end so a future change that accidentally strips
sprite_name on stop or forgets to look at metadata on launch fails
loudly here, not at runtime.
Pair the static-content assertion with a behaviour test that runs the
postlude as a real shell script. Three scenarios exercise the three
branches inside the snippet: remote tip exists (track origin), no
remote (fresh local branch), and already on the target branch
(idempotency). This catches regressions where the snippet still parses
as bash but no longer does the right thing — wrong refspec, missing
-B, swallowed exit code, etc.
Pull the inline docker CLI helpers (`dockerInspectExists`, `dockerStop`,
`dockerCurrentBranch`, etc.) and the generic session-state pollers
(`waitForLatestSessionDone`, `waitForSessionDone`,
`waitForSessionEnvironment`) into `e2e/helpers/docker.ts` and
`e2e/helpers/session.ts`. The spec drops from 763 to 647 lines and the
helpers can be reused by future docker / session tests.
Slice the 471-line button file into three modules with single
concerns: the popover shell stays in `executor-settings-button.tsx`
(168 lines, just trigger + popover layout + icon/status dot),
`executor-environment-info.tsx` owns the env detail rows + status
badge, and `hooks/domains/session/use-task-environment.ts` owns the
fetch/poll/reset lifecycle alongside other session hooks.
Move the popover's prepare-summary reducer (and the
`isFallbackNoticeStep` helper duplicated word-for-word in
`prepare-progress.tsx`) into `apps/web/lib/prepare/summarize.ts`. The
hook moves to `hooks/domains/session/use-prepare-summary.ts` to match
the project's domain-organized hook layout.
The popover's PrepareSummary type and the chat panel's EffectiveStatus
serve different UIs, so the panel keeps its own `deriveStatus` — only
the literal duplicate (`isFallbackNoticeStep`) is shared.
The two env-reuse helpers (`applyExistingEnvironment` and
`applyExistingEnvironmentRuntimeMetadata`) were always called together
and shared the same intent: carry the previous TaskEnvironment forward
into the launch request. Merge them into a single
`reuseExistingEnvironment(ctx, req, env)` so the executor-type gate
runs once for both env-level fields and runtime metadata, and callers
have a single entry point.
Document the layered metadata rule (running wins for container_id;
env wins for sprite_name and other stable IDs) inline so the next
reader doesn't have to reverse-engineer it from two functions.
ExpandSessionDir kept a `{home}/<dotdir>` fallback for the case where
kandevHomeDir was unresolved. In production, ResolvedHomeDir() always
returns a non-empty path, so the fallback only ever fired in tests
that hadn't been updated to pass an instance ID. Returning the empty
string in that case makes container.go skip the bind mount — the
correct behaviour, since without a kandev home there's no session dir
to mount.
`instance_port` is only used as a zap log field — it's never written to
the executor metadata map, so carrying it across resume in
`persistentMetadataKeys` was a no-op.
Strip the inline `git checkout -B` block from the Docker and Sprites
default prepare scripts. The kandev-managed postlude
(KandevBranchCheckoutPostlude) is appended unconditionally after every
user prepare script, so keeping the same logic inline made the default
and the postlude two sources of truth that could drift.
Tier 1 of the post-review cleanup. Each item is independently safe; bundling
because they're all small, mechanical, and lint-passing.
- fix(clone): unify injectTokenIntoURL across Docker and Sprites — Sprites
used to ignore GH_TOKEN entirely (a user with only GH_TOKEN set got 401
on Sprites clone but worked on Docker). Promote one
`injectGitHubTokenIntoCloneURL` helper, accept both env vars
(GITHUB_TOKEN wins, matching Docker), and use the documented
`https://x-access-token:T@github.com/` form on both surfaces.
- fix(executor-popover): swap the <span tabIndex role="button"> trigger
for <Button variant="ghost"> so Enter/Space activation, focus rings on
Safari, and disabled semantics come for free.
- test(docker): rewrite TestDockerStopInstance_PreservesSessionDirOnPlainStop
to drive the real DockerExecutor.StopInstance path with ContainerID=""
(skips the docker daemon, exercises the kandev-side cleanup branch).
Previous version only iterated shouldRunExecutorCleanup strings — would
pass even if StopInstance always cleaned up. Now table-driven over
preserve+destructive reasons with on-disk asserts.
- refactor(orchestrator): drop dead `_ = arg` parameters from persist*
helpers (persistResumeState, persistModelSwitchState, persistLaunchState).
Brings signatures from 9→6 args, under the funlen lint cap.
- refactor(orchestrator): collapse byte-identical isContainerizedExecutor
≡ isCloneBasedExecutorType predicate into one.
- refactor(lifecycle): drop dead getMetadataBool, FileUploader.RunCommand,
expandMountSource {home} arm (no production agent ships {home} in
Mounts; codex_acp_test asserts it), dockerTaskBranch one-line wrapper.
The executor section is informational, not a button — readiness state +
sandbox/container/branch fields. Move it to the left top bar (next to
the integrations menu) where it sits alongside the other contextual
info, drop the outline-button styling, and switch from click-to-open to
hover-to-open via HoverCard with 150ms open / 250ms close so the cursor
can travel to the destructive Reset button without flickering.
The trigger is a borderless icon-with-status-dot span; tabIndex keeps
it keyboard-accessible. Reset still works — it opens the existing
confirmation dialog independently of the hover card.
Tests: pointerEnter instead of click for the hover-to-open path.
The READY badge next to the executor name already conveys ready-state;
the dedicated "Environment ready · 12s" row above the executor info
just added vertical noise. Skip rendering for the ready phase (other
phases that surface non-redundant info — preparing, resuming, failed —
still render). Drop top padding on the executor info section now that
nothing sits above it on the happy path.
The previous tighten only fixed the top half — pb under the field rows
plus the Reset-row's full p-2 still left a generous gap below the last
"Sprite ..." line. Same fix applies whether the popover is showing
Docker or Sprites.
EnvironmentInfo: pt-2 pb-1.5 (was py-2) so the bottom field hugs the
divider, and the Reset-row goes from p-2 to px-2 py-1.5 so the button
sits closer to the text above.
The Sprites/Docker info block under the Environment-ready row had too
much vertical breathing room — the header-to-fields gap and the
inter-row spacing piled up and the popover felt airy. Trim the section
to match the snug rhythm of the rest of the popover (px-3 py-2,
1.5-unit gap before the field list, 1-unit gap between rows). Also
unify the Environment-ready row from p-3 to px-3 py-2 so the dividers
between sections sit at the same vertical cadence.
Profiles created in the UI snapshot the *then-current* DefaultPrepareScript
into their stored prepare_script field. Updates to the default — like
shipping the worktree-branch checkout block — never reach those older
profiles, so users keep landing the agent on the chosen base branch
(main) even though the env row carries a feature branch name.
Append a kandev-managed postlude to whatever prepare_script ran (custom
or default), wrapped in a subshell + `|| true` so it's idempotent and
never blocks agentctl from starting. The postlude prefers the remote tip
when origin already has the branch (resume after a destroyed sandbox),
falls back to checking out the existing local branch, and finally
creates one off the current HEAD.
Sprites needed the WorktreeProvider in resolvePrepareScript so the
postlude's `{{worktree.branch}}` actually expands; Docker already had
it.
The Docker e2e gains a regression case that explicitly stores an older
default-style script (clone, no checkout) and asserts the agent still
ends up on the feature branch — locking the contract in regardless of
which default the user happens to have stored.
Every ACP agent's session dir was bind-mounted from ~/.<agent> on the
host into /root/.<agent> in the container. That exposed host state DBs,
session caches, and lock files to the container — codex's rollout DB
caches absolute /Users/... paths, which don't resolve inside Linux and
the agent silently loses prior thread context on resume.
Replace the host bind mount with a kandev-managed per-container session
dir under <kandev-home>/agent-sessions/<instance_id>/. Before the
container starts, selectively copy the auth-relevant files (the same
files Sprites already copies via RemoteAuth.SourceFiles) into that
dir; everything else — state DBs, sessions/, lock files — stays on the
host and the agent rebuilds them inside the container with
container-local Linux paths.
The change is agent-agnostic: any agent that has SessionDirTemplate
+ SessionDirTarget gets the new behavior automatically. Cleanup runs
on destructive stops only (task/session deleted/archived), mirroring
the Sprites preserve-on-stop rule so plain stops let resume re-attach
to the same agent state.
Codex was the only agent that bind-mounted its host home into the
container via an explicit `Mounts: [{home}/.codex → /root/.codex]`
entry, while every other ACP agent uses `SessionConfig.SessionDirTemplate`
+ `SessionDirTarget`. Move codex onto the same surface so the
host-state isolation work in the next commit covers it uniformly.
This change alone does not stop the host-state leak — the bind mount
still resolves to the host's ~/.codex for now — but it removes the
last agent-specific branch in the runtime config so the docker
container manager can apply one rule for all agents.
* refactor(web): extract shared use-session-actions hook from SessionTab
Move WS-level session lifecycle handlers (set primary, stop, resume,
delete) into hooks/domains/session/use-session-actions so the same logic
can power both the dockview tab context menu and upcoming mobile
controls. Dockview-specific panel removal stays local in SessionTab via
the onDeleted callback.
* feat(web): humanize task and session state labels
Mobile task sheet was rendering raw enum values like "IN_PROGRESS".
Add formatTaskStateLabel + formatTaskSessionStateLabel helpers and use
them in the mobile task sheet badge. Session state badge in the new
mobile sessions picker also reuses the formatter.
* feat(web): mobile session/terminal/repo pickers + multi-terminal support
Address parity gaps on mobile by introducing per-pane header pills that
open bottom-sheet pickers, scoped to the relevant view rather than
overloading the left task sheet.
Pickers:
- Sessions pill in chat header → bottom sheet with launch / set-primary
/ stop / resume / delete (with confirm). Pill shows agent label +
idx/total badge.
- Terminals pill in terminal header → bottom sheet with add / close
(with confirm) / switch. Replaces an earlier chip-strip prototype.
- Repos pill on the top bar (multi-repo only) → bottom sheet listing
task repositories; switching jumps to a session on that repo.
Each pill renders with a chevron-down icon + count badge so the
trigger reads as a button, not a status badge.
Multi-terminal:
- New useMobileTerminals hook wraps useTerminals and auto-creates a
first shell when the env reports zero.
- MobileTerminalPane multi-mounts PassthroughTerminal instances keyed
by terminalId so switching preserves scrollback. Active xterm
registers itself in a small mobile-active-terminal sender registry;
the existing key-bar reroutes input through xterm.paste() so Ctrl /
^C / arrows reach whichever terminal is currently visible, not just
the per-session default shell.
- Skip the WebGL renderer addon on mobile (new disableWebgl prop on
PassthroughTerminal). The WebGL atlas miscomputed glyph scaling
after the desktop→mobile responsive switch and rendered text at
several times the intended font size; the canvas renderer is correct
and fast enough for shell use.
Refactors:
- Left sheet reverts to a single-purpose task list (the sectioned-hub
variant was rolled back in favour of per-pane pickers).
- session-task-switcher-sheet-hooks extracted from the sheet to keep
it under the file-size lint cap.
* fix(web): address review feedback on mobile pickers + multi-terminal
Bots flagged a real first-mount bug and a handful of smaller items;
addressed each:
- mobile-terminal-pane: track xterm in state instead of a ref so the
active-sender registration effect re-runs once xterm is ready. With
the ref, the effect ran once on mount, found xtermRef.current === null
while PassthroughTerminal was still initializing (notably under the
ResizeObserver fallback path when the container starts at 0×0), and
silently never registered — leaving the key-bar wired to the default
shell instead of the visible terminal.
- send-shell-input: wrap the activeSender call in try/catch so a stale
registry entry (xterm GC'd mid-frame) falls back to the WS path
instead of dropping the keystroke and silently consuming modifiers.
Added tests for the activeSender routing branch (success, modifier
consumption, fallback on throw).
- mobile-sessions-section: remove the dead `showCloseConfirm ? "Delete" : "Delete"`
ternary and the unused prop on SessionActionsMenu.
- use-mobile-terminals: clear the auto-create guard when addTerminal()
rejects so the user gets a retry instead of an empty terminal list.
- mobile-repo-pill: sort taskRepos by position before falling back to
the first repo, so the pill label matches the picker's ordering when
no session is active yet.
- session-task-switcher-sheet-hooks: catch loadTaskSessionsForTask
rejection in handleSelectTask (was an unhandled rejection that left
the sheet open on backend errors), and stabilize useCallback deps in
handleWorkspaceChange by spreading the SheetNavOptions fields so a
fresh literal each render doesn't defeat memoization. Extracted
selectTaskWithoutPrimarySession to keep useSheetActions under the
function-size lint cap.
* fix(web): preserve task.repositories in sheet-hooks kanban mappers
mapSnapshotToKanban (workspace switch) and handleTaskCreated (newly
created task) only carried the first repository_id, dropping the full
TaskRepository array. The mobile repo picker (useTaskRepoCount +
MobileReposSection) reads task.repositories.length, so it silently
disappeared after a workspace switch and for freshly created
multi-repo tasks. Map the array with the same shape the SSR mapper
uses (id, repository_id, base_branch, checkout_branch, position).
Also address a small nit: wrap addTerminal in an arrow at the New
terminal button so the click MouseEvent isn't passed as the (unused)
first argument.
* fix(web): de-dup mobile terminal auto-create + add state-label tests + repo-tap toast
- use-mobile-terminals: move the auto-create-first-shell guard from
per-instance refs to a module-level Set keyed by environmentId.
MobileTerminalPane mounts useMobileTerminals three times for the
same session (the pane itself, the picker pill via
useActiveTerminalPillLabel, and the picker sheet content via
MobileTerminalsList — Vaul keeps it mounted while closed). With per-
instance guards, all three effects ran on first commit and each
fired createUserShell, racing into up to three shells. The shared
Set ensures only the first effect to observe an empty list creates,
with retry-on-failure preserved (delete + next render re-checks).
- mobile-repos-section: tapping a repo with no sessions now toasts a
hint to open the session picker, instead of silently doing nothing.
- state-labels: add tests covering known mappings, null/undefined
fallbacks, and unknown-state passthrough for both formatters.
* fix(web): single useMobileTerminals via context + tests + drop dead export
Address two follow-up review items:
- Lift useMobileTerminals to a single MobileTerminalsProvider owned by
MobileTerminalPane. The picker pill and the picker sheet's list now
read state from useMobileTerminalsContext instead of calling
useMobileTerminals themselves. The previous setup had three
independent useState-backed lists that drifted out of sync after
addTerminal/removeTerminal — the pane could show a stale active
terminal and the pill count could lag until a server WS push
reconciled them. The module-level guard (added in 9488bb0c) fixed
the creation race; this fix ensures the live list stays consistent
too.
- Drop the dead `subscribeActiveTerminalSender` export from the mobile
active-terminal registry. The keybar reads via getActiveTerminalSender;
notify/subscribe was speculative scaffolding with no production
consumer. Test trimmed accordingly.
- Add use-mobile-terminals.test.ts covering the auto-create logic
(initial create, skip when shells exist, skip while loading,
multiple instances trigger only one creation, retry after failure).
Module-level guard exposes a __resetAutoCreatedEnvironmentsForTest
helper so tests start from a clean slate.
* test(web): capture binary AttachAddon frames in mobile keybar e2e helper
The mobile-terminal-keybar Playwright spec asserted on captured
shell.input WS frames. After this PR, the on-screen mobile terminal is
a PassthroughTerminal with its own AttachAddon WebSocket, and the
keybar routes input through xterm.paste() — bytes go on the binary
terminal WS, not the kandev gateway's shell.input action. The capture
helper missed those frames and every keybar-asserts-on-frame test
failed (Ctrl+L, Shift+Tab, Esc, arrows, symbols, etc.).
Extend attachShellInputCapture to also decode binary terminal frames
(skipping the resize tag 0x01 prefix used by useSendResize) so the
existing assertions on `data` continue to work regardless of which
transport carried them.
* fix(web): keybar control chars + session pill propagation + terminal close ordering
Round-3 review feedback. Several real bugs that the previous rounds'
mocks hadn't caught:
- mobile-terminal-pane: switch from xterm.paste(data) to
xterm.input(data, true). paste() wraps the payload in bracketed-paste
escapes (\\e[200~…\\e[201~) under bash/zsh/fish, so every control
byte from the keybar (^C, arrow keys, Esc, ^D, Tab) was landing in
the shell as literal pasted text instead of as a control signal.
- session-mobile-layout: TaskChatPanel and the agent PassthroughTerminal
were rendering with the URL-derived `sessionId` prop instead of the
store-driven `effectiveSessionId`. The session picker would update
the top bar and terminal pane (which already read effectiveSessionId)
but the chat content and passthrough agent stayed on the old session
until a navigation refresh. Both consumers now follow
effectiveSessionId; the now-unused `sessionId` prop is removed from
MobileChatPanelContent + MobilePanelArea.
- mobile-terminals-section: rework handleConfirmClose. Stop the remote
shell first (and only mutate the picker on success), advance
setRightPanelActiveTab to a remaining terminal when the closed one
was active so terminalTabValue doesn't dangle, and call
releaseAutoCreatedEnvironment when terminals drop to 0 so the pane
recreates a default shell on next render instead of getting stuck
on "Starting terminal…". Extracted as a useTerminalCloseHandler hook
to keep the list under the function-size lint cap.
- session-task-switcher-sheet-hooks: handleTaskCreated now merges
session-derived fields (primarySessionId, sessionCount, etc.) from
the existing kanban entry, with meta.taskSessionId as the priority
source — so an "edit" call doesn't wipe the existing session and a
"create with session" call sets primarySessionId correctly.
Extracted into buildKanbanTaskUpsert + mergeSessionFields +
mapTaskRepositories helpers.
- session-task-switcher-sheet-hooks: mapSnapshotToKanban now carries
optional step fields (allow_manual_move, prompt, is_start_step,
show_in_command_panel, agent_profile_id) so the kanban store keeps
step capabilities after a workspace switch.
- use-mobile-terminals: add releaseAutoCreatedEnvironment(envId) — a
runtime API to clear the module-level guard for one env, used by the
picker when the user closes the last terminal.
- use-mobile-terminals.test: extract SESSION_ID constant.
* fix(web): restore mobile keybar modifier transforms with manual input routing
Round-3 swapped mobile from ShellTerminal → PassthroughTerminal, which
made AttachAddon auto-send xterm.onData straight to the WebSocket.
That bypassed sendShellInput's modifier transform — so OS-keyboard
input on mobile (e.g. tap Ctrl on the key-bar, then press "l") sent
the raw "l" instead of "\\x0c" (form feed). Multiple e2e keybar tests
caught this; my prior xterm.input(data, true) attempt only fixed the
key-bar tap path, not OS-keyboard typing.
Add `manualInputRouting` + `onWsReady` props to PassthroughTerminal
(plumbed through useWebSocketConnection / startReconnectLoop /
connectWebSocket). When `manualInputRouting` is true, AttachAddon is
constructed with `bidirectional: false` so it only handles PTY → screen,
and the consumer is responsible for forwarding xterm.onData itself.
Mobile TerminalSlot now:
- passes manualInputRouting=true and an onWsReady callback that
registers an active-terminal sender writing raw bytes to that
terminal's dedicated WS (replacing the prior xterm.input registry).
- attaches its own xterm.onData listener that calls sendShellInput,
which applies the key-bar's sticky Ctrl/Shift transforms before
forwarding via the registered sender.
Net effect: both key-bar taps AND OS-keyboard typing flow through
sendShellInput on mobile, so modifiers apply uniformly. Desktop
unaffected (default bidirectional AttachAddon path unchanged).
Also extracted teardownWebSocket into ws-reconnect to keep the
use-passthrough-terminal file under the line-count cap after the
new options.
* fix(web): tighten mobile sender registration + Set Primary + launch return
- mobile-sessions-section: drop the borrowed isSessionStoppable guard
on the "Set as Primary" item. The backend session.set_primary action
has no such restriction, so a COMPLETED or FAILED session was being
silently grayed out. Disable only when isPrimary or state isn't
loaded yet.
- mobile-terminal-pane: register the active key-bar sender only while
the dedicated WebSocket is OPEN. Previously the sender went into the
registry as soon as the tab became active, so during connect /
reconnect it would silently no-op instead of letting sendShellInput
fall through to its WS path. Track ws in state, drop it on close,
and gate registration on readyState === OPEN.
- session-task-switcher-sheet-hooks: early-return after a successful
launchSession so we don't redundantly call setActiveTask after
setActiveSession. Mirrors the existing "sessions exist" branch.
* feat(web): per-task color indicator in sidebar (localStorage)
Adds a color marker to tasks in the left task list for better visual
identification. Right-click a task → Color submenu picks one of seven
colors (or None to clear). Color renders as a 3px bar on the left edge
of the row and persists in localStorage.
* fix(web): address review feedback on task color indicator
- Filter storage events to TASK_COLORS_STORAGE_KEY so cross-tab writes
to unrelated keys no longer wake every TaskItem subscriber.
- Cache the parsed color map at module level, invalidated on writes and
cross-tab storage events. Avoids re-parsing localStorage on every
getSnapshot call (N tasks × N renders).
- Drop redundant typeof string check on Object.entries keys.
- Dim the colored bar to opacity-60 when not selected so colored rows
remain distinguishable from colored-and-selected rows at the left edge.
* refactor(web): remove test-only export from task-colors
Drop the __resetTaskColorsCacheForTests escape hatch from the
production module. Tests now invalidate the cache via a synthesized
cross-tab `storage` event, exercising the same listener path the
production code uses for real cross-tab updates.
---------
Co-authored-by: Kandev Agent <agent@kandev.dev>
* refactor(mcp): reframe cross-task agent message wrapper to authorize action
Receivers were refusing peer-agent messages with "the user can talk to
them directly" and stalling. The old wrapper demoted peer messages
("rather than a direct user instruction") and explicitly licensed
refusal ("you may decline, push back"). New wording states the user is
not reading the thread, frames the peer as coordinating on the user's
behalf, and directs the receiver to act with user-prompt authority.
Reply path through message_task_kandev is unchanged.
* test(mcp): pin authorization wording in wrap_agent_message tests
Three reviewers (CodeRabbit, Greptile, Claude) flagged that the
behavioral intent of the new wrapper lives entirely in the body string
yet has no unit-level guard. Add assertions that pin the reply tool
name, the user-prompt-authority phrase, and the absence of the old
refusal-licensing wording, plus a short comment on the deliberate
privilege-escalation trade-off so future readers understand why peer
agents inherit user-prompt authority.
The bottom Resume button on the chat panel was leaving users staring at
nothing for the 5-30s reconnect window: prepare-progress events are
intentionally suppressed on resume (the workspace is already there), so
the popover sat in the idle phase even while session.state was STARTING.
Synthesize a "resuming" phase from session state when no prepare events
are flowing, rendered as a spinner row in the popover with copy
"Resuming session — Reconnecting to the existing environment…". The
button icon also flips to the spinner so the toolbar advertises the
in-flight work without the user having to open the popover.
Extract the prepare-status reducer + section component into their own
file so the main popover module stays under the 600-line ESLint cap.
Adds a Docker e2e test that exercises the user-visible stop-then-resume
flow: launch session, stop it via WS session.stop, launch a fresh
session on the same task, and assert the same task environment +
container ID are reused and the kandev feature branch is preserved.
Also nudge the executor popover padding: when the prepare-status
section is rendered above the Sprites/Docker info block, the cumulative
top+bottom paddings looked too generous around "Sprites sandbox" /
"Local Docker". Tighten both sections vertically without affecting the
horizontal rhythm or the bottom action row.
StopInstance was calling sprite.Destroy() unconditionally, so every
session stop tore down the cloud sandbox. Resume then always fell into
the missing-sandbox fallback, re-cloned the repo, re-ran the prepare
script, and lost the agent's working tree. That's a 30s+ round-trip
for what should be "stop the process, then re-attach later".
Limit Destroy to terminal lifecycle reasons (task/session
deleted/archived), the same set runTerminalCleanupScript already uses.
Plain stops just close the proxy and leave the sandbox alive.
Also update the existing Docker e2e regex: the feature-branch suffix is
now a deterministic 6-hex-char prefix of the task UUID rather than a
random 3-char tail, so resume on legacy env rows lands on the same
branch.
Now that BaseBranch is preserved on resume for clone-based executors,
the missing-sandbox fallback warning detail names the actual branch
("Cloning … on the same branch main") instead of the "(unknown)"
placeholder. Add a regression test so a future change to the resume
metadata propagation can't silently regress that copy.
When a sandbox is destroyed and we provision a fresh one for the same
task, the agent may have previously committed and pushed onto the
feature branch. The previous prepare scripts only ran `git checkout -b
{{worktree.branch}}` which would overwrite that progress with a fresh
branch off the chosen base.
Try `git fetch origin <feature>` first; on success, `git checkout -B`
against the remote tip so the agent picks up where it left off. Falls
back to a fresh local branch when origin has nothing.
Also add the same checkout block to the Sprites prepare script — it was
missing entirely, so every Sprites session was committing onto main.
Sprites tasks committed straight onto the chosen base branch because
SpritesPreparer didn't populate WorktreeBranch and the orchestrator's
"forward env.WorktreeBranch on reuse" check only fired for Docker.
Mirror the Docker preparer's TaskBranchName generation in Sprites,
extract a shared `nonWorktreeTaskBranch` helper that uses a
deterministic suffix derived from the task ID (so legacy env rows that
predate this change resolve to the same branch on every resume), and
let the orchestrator forward env.WorktreeBranch for any clone-based
executor — Docker or Sprites.
Sprites and Docker resumes were sending the launch request with an empty
BaseBranch, so the in-sandbox prepare script's `git clone --branch <X>`
shifted positional args and failed with "repository '/workspace' does
not exist". applyResumeRepoConfig only set BaseBranch inside the
worktree-specific block; widen that to all executors that require a
clone URL.
Local executors are intentionally left untouched: LocalPreparer reads
BaseBranch and forces a `git fetch && git checkout` against the user's
own workspace, which would clobber the "use current state" path.
José Almeida