Task Center previously split unbound tasks into a global scope and tasks needing attention into a separate filter. Keep Current project and All projects, label unbound tasks, and restore legacy global scope preferences as all-projects without changing task data.
Keep Active, Paused, and Finished filters. Active includes idle, queued, running, waiting_approval, failed, and interrupted tasks; task details and project-scoped approvals remain available.
Release note: Task Center now offers Current project or All projects and groups tasks needing attention under Active, while retaining project labels and approval actions.
Keep successful background conversations in project activity until their chat is visible and loaded. Track completion in memory for the current client session; initial history, failures and cancellations do not create completion notices.
Include whole-file companion changes already present in App and shared UI: preserve the workspace beneath Settings, show inline reply-copy feedback, and keep project settings inside the project menu. Include their component tests and documentation; navigation demo source changes remain unstaged.
Validation: 77 focused activity tests, typecheck and lint passed; Electron component fixtures passed 112 checks. Full suite: 4365 passed, 3 failed, 67 skipped (offline inventory timeout and two heartbeat migration failures). Commit-time staged whitespace check passed; tests were not rerun.
Release note: 后台会话完成后保留已完成提示，进入对应聊天后自动清除，避免错过任务结果。
Align remote ACP tool recovery with local runtime behavior when the model responds after a tool failure and completes normally. Preserve original errors and derive recovery state from committed remote events so reconnecting safely replays partial terminal events.
Validation: 461 focused tests, typecheck, lint, and shared Linux Host recovery scenarios passed. Full-suite validation has two pre-existing heartbeat migration fixture failures involving the pinned column.
Release note: 修复远端 Agent 在读取等工具失败后已正常回复，任务仍被误判为失败的问题，并修复断线重连时的工具状态恢复。
Keep the add button after the last tab in the scrollable strip. Show transparent floating scroll arrows on hover without reducing tab space, preserve keyboard and touch access, and reveal activated tabs.
Update tab interaction documentation and regression coverage. Exclude documentation demos from ESLint.
Validation: 66 focused tests, typecheck and lint passed. Full tests reported Agent package and runtime reuse failures before timeout; desktop visual verification remains pending.
Release note: Workbar tabs now keep the add button beside the last tab and reveal floating scroll arrows when hovering over an overflowing tab bar.
Links requesting a new window were silently blocked. Open them in managed browser tabs with the existing session restrictions, clean up script-closed tabs, and follow click-created tabs in subsequent direct and MCP browser tools.
Validation: 232 targeted tests and Windows Electron E2E passed; typecheck and lint passed. The full suite retains six failures outside the popup path.
Release note: Browser links that open a new page now open in a managed tab, and AI browser actions follow the new page after a click.
Persist conversation pins across restarts and show pinned conversations first, including older history. Anchor action menus above the list without shifting neighboring rows, with keyboard dismissal and sidebar-close cleanup.
Add schema 36 migration, focused regression coverage, and feature documentation. File-level staging also includes existing workbar styles and browser contract adjustments in shared files.
Validation: 145 backend tests and 237 App tests passed; typecheck and lint passed. Full test run timed out with Runtime and Agent package failures; native Electron visual validation remains pending.
Release note: 会话菜单新增置顶和取消置顶，重启后仍保留；操作菜单现在悬浮显示，不再挤动下方会话。
Derive application and default terminal titles from the current locale while preserving renamed and unmarked legacy terminal titles. Add locale-switch and layout-restoration regression coverage.
File-level scope also includes existing browser popup selection and close handling in the sidebar, its regression test, and workbar documentation updates.
Validation: 74 focused tests passed; typecheck and lint passed. Full test runs timed out with failures in settings navigation, Agent packaging, and database migration tests. Electron verification remains outstanding.
Release note: Workbar application and new default terminal titles now follow the interface language, including after restoring a layout; custom and legacy terminal names are preserved.
Load lightweight conversation summaries and hydrate full history only when
needed, preserving search, copy, export, queued requests, and remote questions.
Keep unchanged image reconciliation messages stable so list refreshes no
longer resend the entire saved history.
Prepare Desktop 0.13.6 with the approved bilingual notes and aligned feature
matrices. Agent remains 0.13.0; no LoongArch preview or navigation demo edits
are included. Development validation passed 4301 tests with 67 skipped;
the user requested CI validation instead of repeating local release checks.
Release note: 历史会话按需加载，减少切回应用时的重复保存和内存占用，同时保留完整历史、搜索与续聊。
Preserve the published English and Chinese README community QR-code updates. Desktop 0.13.5, Agent 0.13.0, approved bilingual release notes, and the validated product code remain unchanged.
Record the approved bilingual release notes and align Desktop and Agent version metadata. Refresh both feature lists with image tools, PPTX OCR, Runtime process reuse, and the completed/to-implement status convention.
Agent 0.13.0 requires Desktop 0.13.5. Keep Node 24.19.0 and OpenCode 1.18.29 unchanged, preserve published release history, and exclude the LoongArch preview.
Validation: bilingual release notes verified; 4293 tests passed and 67 skipped; typecheck, lint, version consistency, and 63 documentation links passed. No new real model calls, Host reruns, local production build, or packaging. Full UI-driven image invocation and local/remote switching are not covered by the existing validation records.
Image-only PPTX attachments failed with no indexable text even when a local
OCR model was configured. PPTX parsing now recognizes embedded PNG, JPEG
and WebP images through the existing local OCR path, preserves native text
and slide locators, and honors the selected parsing workflow.
OCR workers previously retained model sessions and WASM memory indefinitely
after successful parsing. Release the worker after 60 seconds with no active
or queued requests, reuse it for continuous work, and reload on demand.
Fast text/index modes still skip OCR. This does not render Office layouts,
recognize inherited master images, or change Agent execution permissions.
Validated four real PPTX attachments (40 pages), idle memory release and a
nine-page reimport after model reload on Windows x64 with PP-OCRv6 Tiny.
49 local OCR requests and zero text-model calls. All 4293 tests passed
with 67 skipped; typecheck, lint and the isolated Electron build passed.
Release note: 修复已配置本地 OCR 时图片型 PPTX 仍无法导入的问题，识别结果保留幻灯片定位。OCR 空闲 60 秒后自动释放模型运行内存，连续解析复用模型，下次使用时自动重新加载。
Record YAML application import/export, configurable market URLs, and future ShareServer distribution responsibilities. Define a local model management entry with managed-process and external-service boundaries; both remain planned capabilities.
Specify read-only OpenCode and Continue checklists above conversations, including remote Agent delivery and validation requirements. Record native runtime test evidence and distinguish planned work from implemented behavior.
Add model-controlled image invocation and read-only runtime assignments. Route image tools through Main so accepted operations retain their originating conversation when chat runtimes change, with saved references, cancellation, and manual recovery.
Keep Node hashing outside browser-shared contracts to restore production builds. Validation includes the production build, typecheck, lint, focused regressions, real Host transport checks, and two successful real image requests. Full UI-driven runtime switching acceptance remains documented as pending.
Release note: 开启图片模型的会话调用后，聊天助手可调用图片生成与编辑工具，Runtime 分配自动管理。图片成果保存在原会话，可继续修改，并独立显示生成状态和取消入口。
Multiple projects could spawn duplicate heavyweight runtimes, while settings
changes and remote reconnection could cancel or strand native conversations.
Share compatible local OpenCode and DeepSeek Harness process owners and
Agent-owned ACP connections, retaining workspace routing, per-session modes,
model attribution, question ownership, targeted cancellation and idle cleanup.
Restore remote pending questions and terminal task projections across reconnect
and desktop restart, and preserve accepted ordinary requests during settings
replacement while cancelling expert-team parents with their replaced scheduler.
Extend regression coverage and Agent type checks. Record measured resource and
real-model validation results, including failures and limitations.
Also include the conversation media-generation design drafts (not an implemented
feature), documentation entry links and dedicated Linux test-directory guidance.
Desktop and Agent changes must be delivered together for shared process results.
Validation: typecheck and lint passed; staged content and documentation links
checked. Prior full regression and live-model evidence are recorded in progress
notes; no new full test suite or model calls were run for this commit.
Release note: 多项目会话现在复用兼容的 OpenCode 和 DeepSeek Harness 进程，减少重复启动并保持工作区、工具和取消隔离。修复远端问答在断线、重启及配置更新后的恢复与状态同步问题，Desktop 与 Agent 需配套更新。
React render failures now show a localized reload page, and renderer crashes offer native reload confirmation. Record renderer and load failures in existing diagnostics without automatic retry loops. Reload may lose unsaved input; the intermittent failure cause remains unconfirmed.
Release note: Recover from desktop rendering failures using a visible Reload action instead of being left without a recovery entry point.
Copied local files previously required the attachment picker. The chat
composer now imports one or more supported files directly from paste,
sharing parsing progress, limits, and batch rollback with file selection.
Text and screenshot paste remain available, and failures preserve drafts.
Validation: prior 166 focused tests, typecheck, and lint passed; 28 targeted
tests passed again before commit. The prior full suite had 4214 passes,
67 skips, and 6 window-recovery failures; window tests passed 10/10 alone.
Native clipboard behavior was verified on Windows; macOS/Linux remain unverified.
Release note: 现在可在文件管理器中复制一个或多个支持的本地文件，直接粘贴到聊天输入框添加附件；可查看读取和解析进度，失败时保留草稿与已有附件。
Subscribe the open Magic Notes workspace to committed database changes from desktop and Agent/MCP writes. Preserve selections and unsaved drafts during refresh and external deletion.
File-level commit includes existing runtime edits in shared files; their remaining dependencies are still in the working tree.
Release note: Magic Notes now updates automatically after Agent/MCP edits while preserving your selection and unsaved drafts.
Move the assistant workbar toggle from the floating content corner to the topbar immediately before the theme switch. Preserve toggle behavior and update layout tests and documentation.
Release note: The workbar toggle now stays beside the light/dark theme switch instead of floating over the content area.
Remove the duplicate recovery button beside New project and its extra layout column. Keep the collapsed failure status and the retry action on the affected project inside the dropdown.
Stage complete files as requested, including existing App snapshot regression coverage and remote runtime design updates in those files.
Validation: 29 ProjectSwitcher tests and the targeted App recovery test passed; typecheck and lint passed. Full test run timed out after 60 seconds.
Release note: Remote recovery retries now stay in the project dropdown, keeping the project header stable when a connection fails.
Opening long conversation lists stretched and repositioned the project menu. Give the conversation submenu its own surface and positioning, aligned to the selected project row with independent viewport clamping and scrolling. Use compact navigation when neither side has room.
Validation: 13 component tests, typecheck and lint passed; 108 Electron fixture checks passed across 14 light/dark scenarios. Full test run timed out after 60 seconds.
Release note: Project activity menus now keep the project list stable when opening long conversation submenus.
Prepare bilingual notes and aligned feature lists for the schema-35 history repair, retained tool summaries, Windows DeepSeek workspace paths, and compact project activity navigation.
Agent 0.11.25 is a maintenance release with unchanged protocol and runtime pins. Its compound packages require Desktop 0.13.2. Desktop targets the standard six platforms without a LoongArch preview.
Make activity interaction tests wait for the loaded conversation and enabled send button, reset response mocks between cases, and wait for completed answer rendering. Accept bilingual maintenance-only Agent notes without inventing features or fixes.
Validation: 4,155 tests passed, 66 skipped; typecheck, lint, and release notes verification passed. The affected 20 tests passed three consecutive runs; all 227 App and Agent release-note tests passed. No local production build or packaging was run.
Add 8px below the activity summary to separate it from new conversation. Remove redundant conversation-row arrows while keeping project submenu arrows, and update the UI documentation.
Validation: 11 ProjectActivity tests and typecheck passed. Lint was interrupted before completion.
Release note: 调整项目活动摘要与新建会话的间距，移除会话列表末尾的多余箭头，保留项目子菜单指示。
Show bounded operation summaries for Continue, DeepSeek Harness and direct-model tools, retaining input details across incremental updates. Use one canonical workspace path for DeepSeek Host startup and ACP sessions so equivalent Windows paths reuse a working Runtime.
Compare persisted subagent tool fields by value, release terminal task caches, and avoid rebuilding caches for duplicate remote events. Upgrade to schema 35 to recompact repeated schema-34 tool blocks while preserving events and provenance. Include regression coverage, validation records and temporary-resource test exclusions.
Validation: 4155 tests passed, 66 skipped; typecheck and lint passed. Full desktop real-model validation covered all three tool-summary paths and both Windows path spellings. Storage validation evidence includes local and remote tool workloads and real database upgrade checks.
Release note: Tool records now show paths, commands and search details across runtimes. Fix DeepSeek Windows workspace failures and repeated subagent history growth; schema 35 reclaims redundant history data while preserving execution records, and older clients cannot reopen the upgraded database.
Replace the production activity dialog with anchored project and conversation menus, preserving exact conversation navigation and adding keyboard and narrow-window coverage. Include the interactive demo and design documentation.
The demo hides idle activity and removes conversation arrows; production still retains those two elements. Validation: typecheck, lint and 11 ProjectActivity tests passed. The full suite timed out after 240 seconds with an agent-package offline inventory test failure.
Preserve OpenCode operation titles in main-session tool summaries, including permission waits, failures and interrupted calls. Keep title trimming and length limits aligned with subagent records.
Update tool display documentation and include the activity-menu wording in UI-DESIGN.md. Validation: 132 focused tests passed; main-process typecheck and changed-file lint passed. Full-suite checks were affected by concurrent activity-menu changes and a test timeout.
Release note: OpenCode main-session tool records now show file paths, commands or operation descriptions beside tool names, matching subagent records.
Repeated full subagent progress snapshots could grow the desktop database to
many gigabytes during ordinary long-running tasks. Local and managed SSH
writes now retain lossless progress changes while preserving conversation
snapshots, event identities, remote replay deduplication and transaction rollback.
Upgrade existing history in bounded worker-thread batches, show progress and
retry controls, and reclaim freed disk space after conversion. Schema 34 cannot
be opened by older clients; a pre-upgrade backup is required for client rollback.
Document reusable database migration guidance and the verified large-database,
portable, local Runtime and Linux Host evidence.
Prepare Desktop 0.13.1 with the approved bilingual notes and aligned feature
lists. Agent remains 0.11.24. Final candidate validation and production builds
will run on main CI, as requested, without additional local release testing.
Release note: 修复长任务和子任务造成的数据库异常膨胀；升级时自动整理旧记录并回收空间，保留聊天、执行详情和结果，并支持退出后继续处理。
Native browser pages could cover application dialogs and overlapping menus
or notifications. Suspend only the viewport while covered and restore the
same active page afterward, including nested dialogs and terminal prompts.
Keep knowledge source controls at their full height, constrain long library
cards and titles, and update automatic names when switching remote targets
without overwriting edited or saved names. Restore packaged release-note
loading now that the history exceeds its previous size limit.
Add focused regressions and real Electron UI evidence. Clean up temporary
Harness fixture directories at test completion instead of accumulating them.
Release note: 修复内嵌浏览器遮挡弹窗和浮层、知识库来源选项被压缩及长名称显示异常；切换远端知识库时自动名称正确更新，并恢复更新说明显示。
Record the approved bilingual release notes against Desktop v0.12.12 and
Agent agent-v0.11.23, rather than listing intermediate development fixes.
Desktop adds external knowledge connections and cross-project activity;
remote native questions remain classified as fixes in both release notes.
Align both feature lists, compatibility guidance and validation evidence.
Upgrade Desktop to 0.13.0 before Agent 0.11.24. Keep OpenCode at 1.18.29,
retain existing published notes, and exclude the LoongArch preview.
Allow fixes-only Agent notes while checking bilingual section consistency.
Validation: 4089 tests passed with 66 skipped; release-note verification,
typecheck, lint and documentation link checks passed. Full remote-provider
acceptance and candidate CI/native release validation remain outstanding.
No release tags or publication are part of this commit.
Completed parallel conversations could leave buffered event connections open.
Chat and native context compaction now cancel their own subscriptions before
ending SDK iteration, without cancelling another conversation or its Server.
Strengthen cleanup-order regressions and verify real OpenCode parallel runs,
cancellation and compaction. The remote Agent uses ACP stdio and transcript
polling rather than this SDK event stream and needs no corresponding change.
Validation: 99 focused tests passed; the final candidate passed 4089 tests
with 66 skipped, typecheck and lint. No external model calls were made.
The titlebar drag-region defect was fixed three times in a row, each time
only for the dialogs someone reported, and the same batch introduced two more
instances. Repeated full-table lookups and production reuse of a verification
helper showed up in several places at once for the same reason.
Add docs/quality/recurring-defects.md describing each type that has actually
recurred here, the shared mechanism or automated test that now covers it, and
the history that justifies the entry. Require reading it from AGENTS.md before
touching overlays, store lookups, retrieval entry points, cancellation, or
render-time derived values, and require extending the shared mechanism rather
than patching one reported case.
Correct the external knowledge technical design where it described the removed
configuration-snapshot checks, the shared retrieval entry point and the
unreachable probe status, and record the refactor and its validation in the
knowledge base progress log.
Remove code with no production path: the probeStatus value 'healthy' that no
probe can produce, the external.transportConfirm string and the
remote-missing and config-invalid state strings that nothing renders,
KnowledgeService.searchHybridMany which lost its last caller when chat search
moved to retrieveMany, and the sourceUrl field that no provider parser
populates.
Keep sourceUrl in externalKnowledgeLocatorSchema because that schema parses
already-persisted citations. Point the affected fixtures at catalog-ready,
the status testInstance actually writes.
Document why ProjectActivity adjusts state during render: the guarded call is
the recommended pattern for resetting state on prop changes here, and the
repository lint rule rejects the useEffect alternative.
The drag-region fix applied to six dialogs left seven overlays uncovered,
including two added in the same batch, so their controls still stopped
responding to hover and clicks where they overlapped the native titlebar.
Three overlays used inset: 38px 0 0 to dodge the titlebar instead of
releasing the drag region, leaving an unmasked strip.
Replace the per-dialog declarations with one shared selector covering all
fifteen overlays and give each one inset: 0. Add a regression test that
discovers every fixed overlay from the stylesheet and fails when a new one
misses the shared rule, so this stops being fixed one dialog at a time.
Release note: 修复 MCP 编辑器、发布说明、远端目录选择、HTML 预览、设置、跨项目活动和外部
知识库实例等弹窗与标题栏重叠时，控件无法悬停和点击的问题。
Binding checks loaded and parsed every external knowledge binding to answer a
single-row question. requireLibrary did this on more than thirty call sites
including one inside a task reconciliation loop, and the knowledge snapshot
scanned the table twice per library across its seven callers.
Add keyed store lookups (getBinding, hasBinding, getBindingsForInstance) that
use the existing primary key, build the snapshot's bindings once into a Map,
and count bindings per instance in a single pass.
Memoize the conversation activity index in the renderer. It was rebuilt on
every render, including once per streaming token, which discarded half the
benefit of the memoized activity summary directly above it.
Release note: 知识库页面、对话预检索和跨项目活动列表在绑定较多时响应更快，流式回答期间的
界面开销降低。
Conversation retrieval called the binding verification helper, so every
question sent the user query as a test query and RAGFlow bindings with graph
or knowledge compilation enabled spent an extra detail request before
retrieving. A failed capability check could also surface
EXTERNAL_KB_CAPABILITY_UNAVAILABLE during a normal conversation.
Retrieval and verification are now separate service entry points sharing one
snippet-bounding helper. The capability pre-check stays in the verification
path and keeps using the binding's configured request timeout.
Replace the five configuration snapshot comparisons and the
EXTERNAL_KB_CONFIG_CHANGED code with the existing per-instance cancellation.
A connection test cancelled by a configuration change now rejects without
writing back a probe result, so disabling an instance no longer records it as
a failed probe. Duplicate binding safety relies on a re-check plus the
existing unique index.
Release note: 外部知识库对话检索不再执行仅用于保存校验的能力预检，RAGFlow 图谱绑定每次
提问少一次请求；停用实例时不再把它错误标记为连接失败。
Render feedback, parsing diagnostics, knowledge confirmations, WeChat binding and image preview overlays through body portals so native titlebar drag regions cannot intercept their controls. Preserve themed form styles and modal focus isolation.
Validation: 320 targeted tests, typecheck and lint passed. Native Windows hover and close checks passed for feedback and parsing diagnostics in light and dark themes at two sizes. Full suite timed out without a summary.
Release note: 修复反馈等弹窗与窗口标题栏重叠时，关闭按钮无法正常悬停和点击的问题。
Cancelling a managed SSH prompt while answering a question could leave the conversation bound to its previous operation and reject the next message. Record the Agent cancellation outcome and reconcile and close the desktop binding before the next prompt.
Add two cancellation regressions and document real SSH/Electron question validation across eight theme and viewport combinations. Local queues and remote question routing are already included in 3b560c9.
Validation: prior focused checks passed 275 tests with 5 skipped; typecheck and lint passed. The final full run passed 4087 tests with 66 skipped and one packaging timeout; that packaging file separately passed 22 tests with 1 skipped. Staged diff whitespace check passed.
Release note: After cancelling a pending OpenCode question on a managed SSH task, you can send another message in the same conversation without an operation identity error.
Render project creation and settings dialogs in a body portal so the Electron title-bar drag region does not intercept their close controls. Cover dialog placement, scrolling, and focus restoration, and document the layout rule.
Release note: Project creation and settings dialogs keep their close buttons clickable after scrolling and restore focus to their trigger when closed.
Connect Dify, FastGPT and RAGFlow through read-only bindings, encrypted local credentials, shared retrieval and persistent citations. Configuration and removal affect GoodBuddy only.
Show active conversations across projects, retain older live work, and improve project dialog scrolling. Queue concurrent questions and route managed SSH OpenCode questions through the Agent. Preserve recoverable read/search errors and partial search results.
Document the Electron UI automation standard, real model and external knowledge validation, and remaining acceptance limits.
Validation: earlier full regression passed 4081 tests with 66 skipped; subsequent focused remote-question checks and real Linux evidence are recorded in feature progress. Precommit typecheck and lint passed; latest project dialog and shared UI suites passed 58 tests. The full suite was not rerun for the final dialog changes.
Release note: Use external knowledge in conversations, find active work across projects, and answer concurrent local or managed SSH OpenCode questions. Managed SSH questions require the updated Agent adapter.
Allow users to close leased browser tabs without cancelling the AI request or revoking its other tools. Closed-page operations report a recoverable error; explicit navigation creates a dedicated replacement tab for direct-model and MCP calls.
Add regression coverage and real Electron lifecycle checks, and update browser behavior documentation.
Release note: 浏览器标签页被 AI 请求占用时也可关闭，其他任务继续运行；再次导航会打开独立新页。
Consolidate composer settings into a compact popover while preserving runtime, mode, sending, and queue controls. Place the settings icon after voice input, limit the panel to 300px, and show knowledge expansion state.
Include the navigation demo, design guidance, localized labels, and composer regression coverage.
Release note: Simplify the conversation input toolbar with compact settings, clearer knowledge expansion, and visible active selections.
Show a clear button when the conversation search contains text. Clearing restores the scoped conversation list and returns focus to the input, with regression coverage for visibility and filtering.
Release note: Conversation search now includes an X button to clear keywords and return to the current project's conversations.
Require a committed page URL before enabling reload. Clear sidebar action errors when leaving or closing a panel, and ignore late failures from its previous scope.
Add regression coverage and document browser error lifecycle and validation results.
Release note: 修复空白浏览器页刷新时报错、关闭浏览器后错误提示仍残留的问题；切换或关闭面板后，旧操作的延迟失败不再重新显示。
Replace full-width tool cards with compact disclosure rows in conversations and subagent progress. Remove redundant runtime summaries, keep expansion under user control, and show successful calls with green checks.
Prioritize errors and results, fold inputs, and provide original-text copying through application notifications. Detail text wraps in bounded scroll areas with no extra display controls.
Validation: 3927 tests passed, 66 skipped; typecheck and lint passed.
Release note: Tool activity is now more compact, with clearer status icons and less repeated text. Expand a record to read and copy results, errors, or input parameters.
Bump the desktop patch version and package matching Chinese and English notes
for image follow-up context, retained questions and answers, simpler Runtime
selection, and docked Git history with automatic project detection.
Review and align both feature matrices with the verified release range.
Agent stays at 0.11.23; no Agent release or LoongArch preview is included.
External knowledge probe tooling remains development-only, not a shipped
external knowledge-base feature.
Validation: 3911 tests passed, 66 skipped; typecheck, lint, and bilingual
release-note verification passed. Production builds use main-branch CI before
the immutable desktop tag; no local production package was built.
mesalogo