Commit Graph
247 Commits
Author SHA1 Message Date
DeepFal 051eda6166 fix(ai): keep surrogate pairs whole when truncating model context (#4088)
* fix(ai): keep surrogate pairs whole when truncating model context

Head/tail truncation sliced strings by UTF-16 code unit, so a cut could
split an emoji into a lone surrogate. JSON.stringify then emits a
`\udXXX` escape that strict providers reject ("lone leading surrogate in
hex escape"), and because the same cut is reapplied every step, the chat
stays stuck on that 400 for every later request.

- Add textBoundaries helpers (sliceHead, sliceTail, safeSliceBounds) and
  use them at every truncation site on the model request path.
- Wrap provider models with a middleware that replaces any remaining lone
  surrogate with U+FFFD before the request body is serialized.

* fix(ai): keep own __proto__ keys when normalizing prompt values

toWellFormedValue cloned changed objects by plain assignment, so an own
__proto__ key (e.g. from JSON.parse) hit the inherited setter: the field
was dropped from the request and the clone's prototype changed. Clone
with the source prototype and define own properties instead.
2026-11-05 22:44:28 +08:00
陈大猫andClaude Opus 5.5 5098f53944 perf(renderer): cut the startup bundle by a third (#4120)
* perf(renderer): cut the startup bundle by a third

Every window parsed ~6.1 MB of JavaScript before the first render. The entry
chunk carried all five UI locales, and the whole AI SDK chunk was preloaded
even though the AI panel is lazy.

- i18n: only English stays in the entry; other dictionaries load on demand
  (plus their same-base fallback, e.g. zh-TW -> zh-CN). index.tsx awaits the
  saved language before the first render, and I18nProvider re-renders when a
  runtime switch finishes loading.
- AI runtime: aiStateSnapshots cleaned up chats through getAgentRuntime(),
  and the plugin AI/agent hosts created their runtimes on mount, which pulled
  the Catty harness and @ai-sdk providers into startup. Cleanup now goes
  through a light handle that the runtime registers (the reversible privacy
  flag and main-process privacy mapping are still cleared without a runtime),
  and both plugin hosts create their runtime on the first request, re-checking
  liveness after the async load.
- Chunking: manualChunks' object form also moved Vite's dynamic-import
  preload helper into vendor-ai (the AI SDK itself uses import()), so the
  entry imported it from there and vendor-ai (649 KB) was preloaded on every
  launch. Vendor chunks now use the function form and the helper gets its own
  1 KB chunk.
- jsqr is loaded when a QR image is imported instead of with the host editor.
- startupBundle.architecture.test.ts guards these invariants.

Startup JS (index.html): 5.29 MB entry + 649 KB vendor-ai + 138 KB radix ->
3.80 MB entry + 1 KB helper + 174 KB radix (-35%). Startup harness, 2000 hosts,
10 runs, warm p50: loadURL -> renderer entry evaluated 321 -> 207 ms,
loadURL -> renderer ready 481 -> 323 ms.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(i18n): preload locale messages before static rendering

* test(i18n): preload locales in remaining zh-CN UI tests

Non-English dictionaries now load lazily, so I18nProvider can initially render English before its effect finishes loading zh-CN. Explicitly await loadLocaleMessages before rendering these localized UI tests to fix the remaining CI failures without timing-based waits or production changes.

* chore: retrigger CI after locale preload fix

Previous tip 38108ce66 was on the branch but the PR head stayed
stuck on d49ad024e after a GitHub 500 during push. Empty commit
to refresh the PR head and re-queue checks.

* chore: retrigger CI after plugin-contract runner cancel

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 03:19:21 -08:00
8b5b0c4416 fix(ai): keep dropped AI chats recoverable with dangling marking + /rebind; reuse workspace id on merge (#4019)
* fix(ai): keep dropped AI chats recoverable with dangling marking and /rebind

Closing a member terminal or dissolving a workspace used to clear the
chat's active pointer, making it invisible (scope rank 0) and
unrecoverable from the UI even though the record survived in storage.

- orphan cleanup now marks dropped chats `scopeDangling: true` and keeps
  record + active pointer intact; the flag clears when the target
  returns or the chat is rebound
- `/chats` lists dangling chats with a numbered inline affordance and
  hint line
- new `/rebind <n>` panel action rewrites `scope` to the panel's current
  scope, clears the marker and restores the active pointer
- i18n keys added for en / zh-CN / zh-TW / es / ru

Stacked on #4018. Refs #4015 (R2).

* fix(workspace): reuse existing workspace id when merging tabs with a workspace

createWorkspaceFromSessions silently no-opped when either merged tab
already belonged to a workspace (e.g. dragging a workspace pane onto a
standalone terminal), and the older dissolve->recreate behavior
regenerated the id that workspace-scoped AI chats bind to.

When one of the merged tabs has a workspace, the orphan tab now joins
that existing workspace (pane inserted next to the anchor session, tab
slot collapsed), so workspace-scoped chat bindings survive. Merging two
standalone tabs still creates a fresh workspace.

Stacked on #4018. Refs #4015 (R3).

* fix(ai): persist cleaned active map on dangling chat rebind

rebindSessionScope was updating React state to drop stale old-scope
pointers, then rebuilding localStorage from the pre-rebind snapshot and
resurrecting them. Persist the same cleaned map via rebindActiveSessionMap.

* fix(ai): persist rebind map inside setState; wire backup actions on terminal panel

- Write cleaned active-session map from inside the rebind updater so a
  deferred React setState cannot skip snapshot/localStorage persistence.
- Pass exportSessionsBackup/importSessionsBackup through TerminalLayerSupport
  and include them in AI_CHAT_SIDE_PANEL_AI_STATE_KEYS so history-drawer
  backup buttons work on the normal terminal path.

* fix(ai): guard backup restore races and use app-wide live targets for /chats

- Skip restore when in-memory sessions already exist; apply snapshot,
  storage write, and AI state emit atomically inside the empty-state path
  so peer useAIState instances see the restore.
- Thread validAIScopeTargetIds into AIChatSidePanel as liveScopeTargetIds
  so /chats does not mark chats bound to other open terminals as dangling.

* fix(ai): open rebound chat in panel view and refresh /chats listing

- rebindSessionScope also sets the panel view to the rebound session so an
  existing draft/other chat does not keep hiding the recovery.
- After /rebind, rewrite the stored listing instead of rebuilding from the
  stale render so the chat is no longer offered as dangling.

* fix(workspace): invert join split when dragged pane owns workspace; refresh /chats titles

- When merging a workspace-owned dragged pane onto a standalone tab, the
  orphan is inserted — invert the drop position so the previewed side matches.
- Compare workspaces in AIChatSidePanel memo so /chats picks up renames.

* fix(ai): validate live sessions array on boot; mark imported orphans dangling

- Reject parseable non-array store values (e.g. {}) via readLiveSessionsArray
  so boot restore is not blocked by a corrupt-but-truthy value.
- After manual import, re-run orphan cleanup with the last known live target
  set so chats bound to closed scopes become dangling and /rebind-able.

* fix(ai): merge imported sessions before orphan cleanup

Apply import merge to the snapshot/ref synchronously, then run dangling
marking, so a deferred setState cannot let cleanup rewrite over the import.

* fix(ai): mark restored/imported orphans dangling even with no live targets

Restore and import do not change live terminals/workspaces, so the orphan
cleanup effect will not re-run. Always call cleanup against the last known
live target set after those paths — an empty set is a valid observed state
(nothing mounted), not an unavailable snapshot — so chats bound to missing
scopes become dangling and /chats can offer /rebind.

* fix(ai): apply backup restore synchronously before dangling cleanup

Do not gate restore + orphan marking on a setState updater having run.
A deferred updater left didRestore false, skipped cleanup, and could leave
restored chats with missing scopes unmarked and unrebindable from /chats.

* fix(ai): complete /rebind from a partial slash picker selection

Selecting Rebind after typing `/re` left the composer unchanged, so a
following number became `/re 1` and could be sent to the agent. Replace the
slash query with `/rebind ` the same way Rename does.

* fix(ai): replace hostIds on rebind; skip non-array backup generations

- Rebind builds a fresh destination scope (including hostIds from the
  panel) instead of spreading the old scope, so history matching follows
  the new terminal/workspace host.
- readLatestGeneration treats valid JSON that is not a session array as
  unusable and continues to older generations, matching corrupt skips.

* fix(ai): revalidate dangling chat before /rebind from stale /chats listing

Skip rebind when the cached entry is deleted or already bound to a live
scope; drop stale listing entries so a second panel cannot steal the chat.

* fix(ai): reject malformed /rebind numbers

Require the entire trimmed argument to be digits so /rebind 1foo and
/rebind 1.5 show usage instead of silently targeting index 1.

---------

Co-authored-by: binaricat <binaricat@users.noreply.github.com>
Co-authored-by: 陈大猫 <16399091+binaricat@users.noreply.github.com>
2026-10-27 20:26:21 -07:00
GOUKI9999andNetcatty Maintainer afb0585016 feat(ai): /rename & /chats slash commands + AI session durability layer (#4018)
* feat(ai): add /rename and /chats slash commands plus a session durability layer

Panel slash commands (#4015 feedback):
- /rename <title> renames the current chat via updateSessionTitle; the
  command is intercepted in ChatInput's system-command branch and never
  reaches the agent. Empty title shows a usage hint. The history drawer
  gains an inline rename affordance.
- /chats renders a dismissible system-style note listing the panel's
  visible chats (title / agent backend / workspace-or-terminal scope with
  a dangling marker / relative last-active time; current chat marked).
  Deliberately not named /resume: session.resume already means
  flow-recovery-after-cancel. Listing logic is pure (aiChatsListing.ts)
  so a future rebind action can attach per-entry actions.
- i18n keys added for en/es/ru/zh-CN/zh-TW.

Durability layer (additive; cleanup/eviction semantics untouched):
- Every successful store write schedules a debounced (~2s) mirror of the
  exact persisted JSON to {userData}/ai-sessions-backup/ through a new
  aiSessionsBackupBridge: atomic temp+rename writes, 5 rolling
  generations, timestamped names.
- On boot, a missing/empty live store is restored from the newest
  parseable backup generation. The asymmetry is deliberate: empty ->
  restore, non-empty -> leave alone, so restored data can never
  resurrect explicitly deleted chats.
- Manual export/import in the history drawer via real file dialogs
  (fixed-path fallback where dialogs are unavailable); import merges by
  id with newer updatedAt winning.

Tests: bridge rotation/corruption/rejection, restore decisions, merge
semantics, debounce throttle, mirror-on-write, slash interception and
/chats listing seams. eslint clean; tsc unchanged vs baseline.

* fix(ai): repair sessions backup IPC and restore/import guards

- Forward write/export payloads through preload unchanged so the main
  bridge receives a string json (was double-wrapped → invalid-json).
- Restore from disk only when the live store is missing or corrupt;
  treat a valid [] as an intentional clear. Flush the debounced mirror
  on beforeunload so deletions land before quit.
- Require id + messages[] + finite updatedAt when parsing backup/
  import entries so malformed objects cannot reach pruneSessionsForStorage.

Co-authored-by: Netcatty Maintainer <binaricat@users.noreply.github.com>

* fix(ai): complete /rename from partial slash and harden import shape

- Selecting Rename from the slash picker replaces the active slash token
  with `/rename ` so a following title is intercepted, not sent to the agent.
- parseBackupSessions also requires title, agentId, and a scope object with
  a string type before accepting an imported/backup entry.

Co-authored-by: Netcatty Maintainer <binaricat@users.noreply.github.com>

* fix(ai): validate backup messages and skip prune on synced restore

- Drop backup/import sessions whose messages lack id/role/timestamp.
- When cloud AI-session sync is enabled, restore parsed sessions without
  pruneSessionsForStorage so local retention cannot truncate synced history.

Co-authored-by: Netcatty Maintainer <binaricat@users.noreply.github.com>

* fix(ai): export prepareRestoredSessions helper

Missed in the previous commit; required by useAIState boot restore and tests.

Co-authored-by: Netcatty Maintainer <binaricat@users.noreply.github.com>

* fix(ai): require message content on backup import; raise disk cap to 64MB

* fix(ai): normalize corrupt sessions store to [] before boot restore

* fix(ai): skip non-array backup generations when reading latest

* fix(ai): emit sessions change on backup restore; skip unusable gens

* fix(ai): treat empty backup gen as wipe; align gen usability with parser

* fix(ai): seed backup on boot for live store; report durable import count

* fix(ai): emit sessions change after backup import

* docs(ai): clarify sessions backup recovers in-profile store only

---------

Co-authored-by: binaricat <binaricat@users.noreply.github.com>
2026-10-27 20:04:42 -07:00
GOUKI9999 02f8c798f7 feat(ai): hydrate model pickers from last-known-good and fetch only on demand
The list-models pipeline already served LKG on failure; this wires the
rest of the LKG-first, manual-refresh policy end to end.

Main: the list-models IPC accepts a force flag — manual refresh always
skips the in-memory cache and the persisted entry and goes live, and a
forced success overwrites the persisted catalog.

Renderer: the panel passes force only from the retry banner and the
picker's refresh button; panel-visibility and send-preflight loads stay
non-force and replay the renderer cache, so with a persisted catalog
main answers them without spending the live-fetch budget.

Picker: a footer refresh icon-button next to the status chip, visible
whenever any fetched catalog (live or last-known-good) or the degraded
warning is present; the spinner row remains the in-flight feedback.
Tooltip localized as ai.chat.refreshModelList in en/es/ru/zh-CN/zh-TW.

Fresh installs are unchanged (no persisted entry → live fetch as
before); existing users stop seeing jitter warnings on panel open and
pick up new CLI models via the refresh button.
2026-10-23 17:22:25 +08:00
陈大猫 5ca36721b5 fix(ai): show MCP tool names on External MCP approval alerts (#3779)
Follow-up to #3771: notification/card label maps internal rpc ids (e.g. netcatty/exec) to catalog MCP tool names (terminal_execute).

Real-app smoke PASS on 80c28d8e.
2026-10-20 12:41:41 +08:00
b2503ff6ba feat(ai): alert when an External MCP approval is waiting (#3771)
* feat: alert when an external MCP approval is waiting

Show a system notification and request taskbar or Dock attention so confirm-mode approvals are not missed before the existing timeout.

Co-authored-by: 陈大猫 <binaricat@users.noreply.github.com>

* fix: stop MCP approval attention when a later window focuses

Watch app browser-window-focus while a flash or dock bounce is active so a
window created after the alert starts can clear it.

Co-authored-by: 陈大猫 <binaricat@users.noreply.github.com>

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: 陈大猫 <binaricat@users.noreply.github.com>
2026-10-20 11:54:44 +08:00
陈大猫 2ce0164fa9 Merge pull request #3642 from GOUKI9999/feat/unified-model-catalog-ux
feat(ai): unify SDK model catalog loading and sources
2026-10-15 22:48:27 +08:00
bincxz 239804bf62 fix(ai): retain native Copilot lifecycle and clean undated catalog labels 2026-10-15 21:01:19 +08:00
bincxz dfc7961b68 fix(sync): complete portable chat and preservation flows 2026-10-15 18:26:17 +08:00
bincxz d9632ab985 Merge latest model catalog fixes from PR 3624 2026-10-15 18:14:00 +08:00
0ed481c397 feat(ai-chat): fork a conversation from a completed response (#3604)
* fix(#3600): automated AI fix

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix: address Codex review on PR #3604

* fix(ai): narrow chat forks to independent transcript copies

* test(ai): keep fork render check correctly typed

* fix(ai): remove forked output arguments and support keyboard access

* fix(ai): guard signed fork reasoning and scan eligibility once

* fix(ai): remove bare source output references from fork text

* fix(ai): sanitize fork attachment previews without changing data

---------

Co-authored-by: netcatty-bot <308658023+netcatty-bot@users.noreply.github.com>
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com>
2026-10-15 14:38:53 +08:00
GOUKI9999andClaude Sonnet 5.5 a76971e2c5 fix(ai): keep Grok effort-decorated current id, ASCII-only sources; merge #3624 fixes
- last-known-good serving keeps grok-4.6/high when the grok-4.6 row exists.
- Replace non-ASCII dashes/middle dot in files that were ASCII-only.
- Merge #3624 (suffixed Copilot row, retry warning on retained catalog).

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SX3d2XMHxBEpU7VqpvXbdm
2026-10-14 15:13:39 +08:00
GOUKI9999 0006634f6b fix(ai): show build-time presets while the opencode catalog loads
opencode was the only SDK agent whose picker rendered nothing until its
live catalog pull resolved: the #3584 suppression replaced presets with
just the user's custom ids while runtimeModelLoading was set. Align it
with every other backend — presets stay selectable during loading, the
live catalog replaces them on arrival, and #3585's retry/warmer
mechanics are untouched. Adds a regression guard against reintroducing
the opencode gating.
2026-10-12 18:12:34 +08:00
GOUKI9999 e642c4953f feat(ai): unify model catalog UX across SDK agents
Implements the unified model catalog UX on top of the last-known-good
persistence layer (#3624):

- R1 source provenance end-to-end: list-models now returns
  source: 'live' | 'last-known-good' plus fetchedAt (ISO) on live
  successes and last-known-good serve-outs; the pre-flight cache replays
  the original provenance on hits. The renderer cache persists
  source/fetchedAt alongside the catalog, and the build-time preset
  fallback derives 'presets'.
- R2 loading state for all backends: drop the opencode-only gate on
  modelCatalogLoading so every SDK backend shows the composer loading
  state while its live fetch is in flight (opencode behavior unchanged).
- R3 unified status chip: ComposerModelPicker renders a muted footer chip
  for every backend - live / cached with relative last-success time /
  built-in default list - replaced by the spinner row while loading.
  Keys added to en/es/zh-CN/zh-TW (ru keeps its pre-existing ai.opencode
  gap).
- R4 copilot headless wrapper: generated per-resolve wrapper (baked with
  the platform exe the npm loader would spawn) is handed to
  RuntimeConnection.forStdio, spawning with windowsHide:true so no
  console window flashes in Electron. Falls back to the loader when the
  exe cannot be resolved (window may flash - documented degradation).
2026-10-12 16:53:00 +08:00
Netcatty-Botandnetcatty-bot 1209bfa61c fix(ai): opencode model catalog loading state, retry, and warmer list server (#3585)
* fix(#3584): automated AI fix

* fix: address Codex review on PR #3585

* fix: address Codex review on PR #3585

---------

Co-authored-by: netcatty-bot <308658023+netcatty-bot@users.noreply.github.com>
2026-10-08 15:16:49 +08:00
陈大猫 f91b2cd04e feat(ai): scope external MCP approvals to connection or host (#3569)
* feat(ai): scope external MCP approvals to connection or host

* test(ai): keep MCP fixture in Netcatty temp directory

* fix(ai): refine external MCP approval UI and grant safety

* fix(ai): wrap long approval targets and support Node 22

* fix(ai): keep automatic scripts in observer mode

* fix(sftp): release preflight markers after transfer

* fix(sftp): retain preflight cap through conflict resolution

* Keep SFTP preflight cap across panel remount

* Update worker terminal metadata expectations

* Persist SFTP preflight skip state across restart

* Limit persistent host approvals to saved hosts

* Scope host-open approvals to the selected saved host

* Exclude serial hosts from persistent approval scopes
2026-09-30 09:57:47 +08:00
f1425c8d5a feat(ai): add MiMo Code (mimo) agent support (#3500)
* feat(ai): add MiMo Code (mimo) agent support scaffolding

Registry/wiring only: discovery entry, managed-agent metadata, model presets, icon, i18n and Settings UI for Xiaomi's MiMo Code CLI (an OpenCode fork). The SDK driver is intentionally not wired yet - it depends on the pending @mimo-ai/sdk parity probe, so DRIVER_REGISTRY keeps a TODO and run-turn/list-models reject the mimo backend until then.

* feat(ai): add MiMo Code (mimo) agent support

Adds Xiaomi's MiMo Code CLI as a managed agent backend, mirroring the existing
OpenCode integration (discovery, managed-agent metadata, model presets, icon,
i18n, Settings UI) and adding the SDK driver that was left as a TODO in the
scaffolding commit.

Driver notes:
- New electron/bridges/aiBridge/sdk/mimoDriver.cjs reuses the pure translation
  helpers exported by opencodeDriver.cjs (config/prompt building, event
  translation, model mapping) because MiMo Code is an OpenCode fork and speaks
  the same wire protocol.
- It does NOT use @mimo-ai/sdk's createOpencode(). On 0.1.15 that helper spawns
  a binary named "opencode", waits for an "opencode server listening" banner,
  and passes config via OPENCODE_CONFIG_CONTENT; the mimo CLI prints
  "mimocode server listening on <url>", reads MIMOCODE_CONFIG_CONTENT, and ships
  as "mimo"/"mimo.exe". So createOpencode() can only ever time out. The driver
  runs "mimo serve" itself, parses the readiness URL, and attaches the SDK's
  createOpencodeClient({ baseUrl }).
- Binary resolution honours binPath, MIMOCODE_BIN and MIMOCODE_BIN_PATH, then
  falls back to PATH via resolveCliFromPath (the npm shim is mimo.cmd on
  Windows, which Node cannot spawn directly).
- @mimo-ai/sdk ^0.1.15 added as a dependency.

Verified on Windows against mimo 0.1.15 + @mimo-ai/sdk 0.1.15:
- "mimo serve --port" boots and reports its URL; driver parses it.
- SDK symbols used all exist: createOpencodeClient, client.config.providers(),
  client.global.event() (SSE stream), client.session.create/abort/promptAsync.
- Live provider catalog reads 2 providers (xiaomi, mimo); MIMO_MODEL_PRESETS
  updated to the real lower-case provider-prefixed ids and its TODO removed.
- Server process is killed on close (taskkill tree kill), no orphan processes.

Not yet verified at runtime:
- A full end-to-end turn (no prompt was sent, so no tokens were spent): text /
  reasoning / tool-part streaming and abort handling are untested against a
  live model.
- MIMOCODE_CONFIG_CONTENT accepting the OpenCode config schema (permission,
  mcp, model, share, autoupdate) end-to-end; only that the env var exists in
  the binary has been confirmed.
- Whether the loopback server needs MIMOCODE_SERVER_PASSWORD when the API
  client attaches.
- The placeholder public/ai/agents/mimo.svg still needs an official asset.
- Multi-turn resume (resumeSessionId) and MCP injection against the live CLI.

* feat(ai): inject MIMOCODE_BIN for discovered MiMo Code agents

Mirror the OpenCode OPENCODE_BIN handling so a discovered or re-detected
mimo binary path reaches the driver, and cover it with the same assertion
shape as the opencode managed-agent env test.

* fix(ai): use MiMo branding in the mimo driver

* test(ai): cover the mimo driver

Add a unit test file for the MiMo Code driver, matching the existing
opencode driver test in shape, runner (node:test via node --test) and
`.test.cjs` extension. It is the last driver in the registry without
one.

Coverage:
- spawn/launch: `mimo serve --hostname/--port` argv, binary resolution
  (explicit binPath / MIMOCODE_BIN / MIMOCODE_BIN_PATH), pass-through
  env and MIMOCODE_CONFIG_CONTENT; `child_process.spawn` is stubbed, so
  no real `mimo` process is ever started.
- readiness: parsing of `mimocode server listening on <url>` from
  stdout (including split chunks) and stderr, the OpenCode spelling,
  and the MIMO_SERVE_TIMEOUT_MS timeout path (asserts the rejection and
  that the child is killed).
- abort/cleanup: stopMimoProcess kill/idempotency/swallow, the Windows
  taskkill tree kill, server.close() teardown, and abort mid-turn with
  a throwing close().
- branding: OpenCode -> MiMo Code for the busy/idle/status, failed and
  tool-failure fallback strings, with tool output and non-brand
  technical errors left verbatim.
- happy path: session create -> streamed event translation -> done,
  plus listMimoModels catalog mapping and the empty/aborted paths.

* fix(ai): filter MiMo events by session and stop the server behind the npm shim

Address review on #3500:

- The global SSE stream carries every session on the server, and
  runMimoTurn translated all of them, so another session's text could leak
  into the reply and its session.idle could end the turn early. Skip events
  whose session id is not ours, reusing the OpenCode driver's
  getOpenCodeSessionIdFromEvent (now exported). Covered by a test with
  interleaved sessions.
- On macOS/Linux the official @mimo-ai/cli `mimo` is a Node shim that runs
  the native binary through spawnSync, so killing the shim left the server
  listening. `mimo serve` is now spawned detached on POSIX and
  stopMimoProcess signals its whole process group (SIGTERM, then SIGKILL
  after a grace period). Windows keeps taskkill /T.

Verified against the real @mimo-ai/cli@0.1.15 npm shim: on Linux (WSL
Ubuntu 22.04) the listener is the native .mimocode child; before this change
it was still listening 3s after close(), now the port is released within 1s
and no mimo process remains. On Windows the .cmd shim path also releases the
port within 1s.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: harden MiMo CLI integration before merge

* fix: wait for MiMo events and handle setup cancellation

* fix: finish MiMo stream handling and scope skill access

* fix: preflight MiMo skill paths before one server startup

* fix: scope MiMo skill reads to trusted roots

* fix: allow MiMo skills from ancestor project roots

* fix: honor MiMo observer and confirm command permissions

* fix: clear MiMo approval when its service exits

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com>
2026-09-29 14:51:22 +08:00
fc55e9ac17 feat(vault): let Skills + CLI operate notes with the same approval as MCP (#3512)
* feat(vault): let Skills + CLI operate notes with the same approval as MCP

Expose vault note list, read, write, and markdown import on the tool CLI so agents can save generated documents only after the user approves.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(vault): keep explicit empty note fields in the CLI

A supplied empty string was treated as an omitted flag, so note clears and empty markdown imports never reached the note service.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(vault): allow an explicit empty note body on CLI create

notes create --content "" was rejected as a missing flag, so it did not match vault_notes_create.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(vault): stop echoing imported note bodies and recognize leading CLI flags

A batch import resent every document over IPC, and a leading --json hid the notes command from artifact inference.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(vault): make note CLI import safe and fail writes cleanly

---------

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com>
2026-09-29 12:22:35 +08:00
Netcatty-Botandnetcatty-bot 078353dc87 fix(ai): allow custom model ids in composer picker and persist manual picks (#3535)
* fix(#3534): automated AI fix

* fix: address Codex review on PR #3535

* fix: address Codex review on PR #3535

* fix: address Codex review on PR #3535

* fix: address Codex review on PR #3535

---------

Co-authored-by: netcatty-bot <308658023+netcatty-bot@users.noreply.github.com>
2026-09-27 14:57:10 +08:00
netcatty-bot 0b96d70702 fix: address Codex review on PR #3359 2026-09-12 05:59:48 +00:00
fe2f14f032 feat(ai): add Mention Note context picker for Vault Notes (#3290)
* fix(#3289): automated AI fix

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix: address Codex review on PR #3290

* fix(ai): apply note attachment budgets synchronously

* fix(ai): reuse normalized note search fields while typing

* fix(ai): retain draft state binding in panel host

* fix(ai): reject carriage returns in note references

* fix(ai): protect continuation before pruning note payloads

* refactor(ai): keep note mentions as lightweight references

* fix(ai): gate note references on reader availability

* fix(ai): retain note references ahead of replay prose

* refactor(ai): share one instruction across note references

* fix(ai): keep accepted note references within replay space

* Bound agent note reads and add resumable literal search

* Preserve compact note requests and announce picker selection

* Retain mixed attachment context in compact note replay

* Keep imported long note identities readable

* Keep note picker selection and untitled labels consistent

---------

Co-authored-by: netcatty-bot <308658023+netcatty-bot@users.noreply.github.com>
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com>
2026-09-07 14:32:08 +08:00
陈大猫 39d7c38a6a fix(ai): preserve resumed terminal chats after workspace merge (#3225) 2026-08-31 15:03:01 +08:00
bincxz 0288d1610a fix(ai): skip no-op AI config updates and simplify permission picker
Avoid rewriting thinking/provider maps when values are unchanged so
mounted AI panels do not re-render. Put permission title and short
description on one line and drop the Cursor CLI parenthetical.
2026-08-20 11:52:29 +08:00
bincxz b701df8822 fix(ai): align Catty thinking levels with vendor capabilities
Stop advertising Gemini 3.7 Flash minimal, GPT-5.x-chat reasoning, and
Claude 4.7+ extended thinking. Map GPT-5 Off to minimal, send adaptive
effort on Claude 4.6+, and keep composer Enter on an exact or custom id.
2026-08-20 11:05:12 +08:00
bincxz 97ef84db06 fix(ai): persist discovered context windows from composer picks
Codex P2: switching to a catalog model dropped context_length, so
Catty budgeted 128k instead of the advertised window. Keep the
metadata on picker models and write it to modelContextWindows.
2026-08-19 23:51:05 +08:00
bincxz 8bce8c15a0 fix(ai): do not display unset thinking as Off
Codex P2: models whose menu includes Off still showed Off for users
with no stored preference, while the request sent no override. Leave
the chip unselected until the user picks a level.
2026-08-19 23:40:09 +08:00
bincxz 33dc249d74 fix(ai): do not Enter-select custom IDs on external agents
Codex P2: hiding the custom-model row still left Enter falling back
to the typed query, which send-time validation then discarded. Only
use the query as a custom ID when Catty offers that action.
2026-08-19 23:10:34 +08:00
bincxz b490eb3273 fix(ai): keep unset Catty thinking as provider default
Codex P2: opening the composer on Gemini Pro persisted Low for users
who had never chosen a thinking level. Only remap an explicit stale
value; leave an empty preference unset so the request uses the
provider default.
2026-08-19 23:03:56 +08:00
bincxz 55852918c2 fix(ai): sync composer model prefs across retained panels
Codex P2: pinning or recents in one parked ChatInput only updated that
instance. Subscribe every mounted composer so sibling terminals see
the same recent/pinned list without remounting.
2026-08-19 22:57:21 +08:00
bincxz 6e99c9c1dc fix(ai): drop stale external-agent pinned models
Codex P2: refreshed CLI catalogs still showed saved IDs that the
catalog no longer contains, then send-time validation swapped them.
Only Catty may keep a custom ID that is not in the live list.
2026-08-19 22:45:11 +08:00
bincxz 4eddbfab22 fix(ai): hide CLI custom models and count thinking budget once
Codex P2: external-agent pickers advertised a custom ID that send-time
validation ignored. Offer Use custom model only for Catty providers.

Codex P2: the thinking budget was added to both compaction max output
and reservedTokens, so it was subtracted twice. Keep it only on
compaction max output.
2026-08-19 22:29:33 +08:00
bincxz 73a9669bf7 fix(ai): hide Anthropic thinking on models without extended thinking
Codex P2: Claude 3 Haiku still offered Low/Medium/High and would send
an enabled thinking budget. Gate the chip and the provider option on
Claude 3.7 / 4+ model ids. Also memoize Catty thinking levels.
2026-08-19 22:15:34 +08:00
bincxz 5971e03ffe fix(ai): drop stale thinking levels and hide Gemini Pro Off
Codex P2: switching away from Gemini 3 Flash left `minimal` selected
on models that do not advertise it. Persist a valid level for the new
model.

Codex P2: Gemini 3 Pro and 2.5 Pro cannot disable thinking, so Off
was a lie. Those menus only offer low/medium/high.
2026-08-19 22:10:09 +08:00
bincxz 3123c33914 fix(ai): bust model cache on key change and label Gemini minimal
Codex P2: catalog cache ignored API key rotation, so a replaced
credential could keep the previous account's models. Fingerprint the
stored secret in the cache key.

Codex P2: Gemini 3 Flash cannot turn thinking off; Off was sending
thinkingLevel=minimal. Surface Minimal on the chip instead of Off.
2026-08-19 22:02:03 +08:00
bincxz be63983693 fix(ai): hide Catty thinking chip on non-reasoning models
Codex P2: OpenAI-style models such as gpt-4o rejected reasoning_effort
when the chip always offered Off/Low/Medium/High. Gate the chip and
the provider option on known reasoning model ids, and send explicit
none when Off is selected on a reasoning model.
2026-08-19 21:45:15 +08:00
bincxz ad16f19268 fix(ai): tighten picker chrome and align provider action buttons
Remove the provider divider and search-field outline in the sidebar
model picker. Size Save/Test/Cancel to the same h-8 box without
changing the global Button styles.
2026-08-19 21:32:18 +08:00
bincxz 3d154ca235 fix(ai): switch providers from a one-line submenu
Keep the composer model menu as a single-column list. Provider
switching now drills in from the header row instead of a left rail.
2026-08-19 17:35:44 +08:00
bincxz 4acf254f72 feat(ai): unify sidebar provider, model, and thinking pickers
Give Catty a two-column provider/model menu in the composer, add search
plus recent/pinned models, and move thinking effort onto its own chip
so Claude, Codex, Grok, and Catty share the same compact controls.

Closes #2137
Closes #2637
Closes #1884
2026-08-19 17:23:06 +08:00
bincxzandCursor d422a27715 Merge origin/main into feature/app-lock-screen
Reconcile the app-lock feature with main's refactored renderer
architecture (publishers/hosts/AppShell) and main-process changes:

- index.tsx/App.tsx: AppLockGate keeps ownership of useSettingsState and
  useAppLockState; new AppLockRuntimePublisher publishes the runtime via
  appRuntimeBridge; SettingsPublisher now binds the gate-owned settings.
- AppSideEffects: defer deep links, tray port-forward toggles, global
  hotkeys and window close commands while locked; flush on unlock.
- main.cjs: keep fresh-session re-lock before main's reusable-window
  fast path; before-quit guard delegates to handleBeforeQuit with main's
  async plugin-shutdown commitQuit and pending-update cancellation.
- globalShortcutBridge: keep locked-state tray redaction on top of
  main's deferred tray panel show logic.
- preload: adopt main's lastPayload replay for terminal popup config
  (superset of the gate-delay replay fix).
- vite config: keep dev CSP plugin, adopt main's warn-only stale
  optimize-dep handling; electron-builder: keep Windows Hello helper
  extraResource with main's CLI-controlled arch targets.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-18 16:30:30 +08:00
bincxz c87f488553 fix(ai): keep history markdown warmup off the first-type window
A 300ms initial delay let ChatMessageList start the Streamdown import
while the user was still opening the panel to type. Restore the 4s
expand grace and drop a pending Shiki plugin setState if the message
unmounts before the idle callback lands.
2026-08-15 03:18:25 +08:00
bincxz a2ac9c966b fix(ai): render Streamdown prose while the composer stays focused
Blocking hydrate on focus left history as raw markdown for as long
as the input was focused. Shiki still waits for a real idle gap.
Prose, lists, and inline code can mount after typing stops even if
the textarea still has focus.
2026-08-15 03:07:11 +08:00
bincxz b02d52516c fix(ai): load Streamdown without Shiki until a code fence is idle
messageResponse pulled @streamdown/code on first markdown import, so
opening a chat with history paid for the full highlighter even when
the reply was just prose.

Render with CJK only first. When a fence exists and the composer is
idle, lazily import the isolated Shiki plugin.
2026-08-15 03:03:38 +08:00
bincxz 21f5516586 fix(ai): do not hydrate Streamdown while the composer is focused
Typing a few characters then pausing still hitch because markdown
hydrate only checked the 800ms activity window. History rows would
mount Streamdown while the textarea still had focus.

Treat focused/composing as busy, wait 2s after the last key, and
keep idle IPC on the same predicate.
2026-08-15 02:56:27 +08:00
bincxz cdcb628887 fix(ai): keep ChatInput chrome still while the composer text changes
The first glyph still re-rendered the whole 1500-line ChatInput because
hasComposerText lived in the same component as the model picker. Token
usage snapshots also broke memo with a new object each time.

Move send-button emptiness onto a tiny external store so only that
control updates, and compare contextUsage by value in ChatInput's memo.
2026-08-15 02:52:22 +08:00
bincxz 38912ac9bd fix(ai): keep the composer until send preflight accepts
Clearing the box immediately after onSend overwrote the pending draft
while handleSend was still awaiting provider sync. A rejected sync or
stale-scope abort then left the prompt gone even though no user
message was created.

Leave the text in place until the parent clears the draft after the
send is accepted. Unmount commit still uses a ref so setter identity
changes cannot write the prompt back.
2026-08-15 02:42:20 +08:00
bincxz 338722cd89 fix(ai): stop per-keystroke draft flushes and ChatInput memo breaks
Typing still hitched because each IME commit started a 120ms
updateDraft, which re-ran the whole useAIState hook. Parent also
passed a new hosts array and providerSwitcher object on every
render, so ChatInput's memo never held. field-sizing-content
relayouted the textarea on every glyph.

Keep typed text in pending refs until send/hide/unmount. Memoize
mention hosts and the provider switcher. Use fixed field sizing
in the composer.
2026-08-15 02:35:50 +08:00
bincxz f7ca60bab4 fix(ai): clear the composer as soon as a message is sent
Send awaits provider sync before dropping the draft. During that
wait, ChatInput's onChange cleanup could fire on a new setter
identity and write the still-visible text back into the pending
draft, so the uncontrolled box kept the last prompt.

Clear the field right after onSend starts, and only commit on true
unmount via a ref so callback identity changes cannot restore it.
2026-08-15 02:30:14 +08:00
bincxz 1f1af89023 fix(ai): reset the uncontrolled composer when a draft is cleared
New Chat wrote an empty composer-only draft, which the text-only
equality treated as identical to the previous typed draft. The panel
skipped re-render, so the uncontrolled textarea kept the old string
and send could still submit it.

Treat emptying draft text as a lifecycle change so listeners wake, and
commit the live box on ChatInput unmount so mid-IME hide still flushes.
2026-08-15 02:17:44 +08:00
bincxz a36980e9d2 fix(ai): keep composer typing off the draft notify path
Every keystroke was re-rendering ChatInput and the first character
created a store draft that broadcast AI state, so the first 1-2 glyphs
painted and the rest dumped after a stall. IME composition hit the
parent on every pinyin letter, and Chromium spellcheck loaded on the
first word.

Keep the textarea uncontrolled, commit parent state on idle/blur/send,
and treat missing-to-composer-only drafts as text-only so store
listeners do not wake. Draft flush, send, and hide still persist the
current box text.
2026-08-15 02:07:03 +08:00