0.4.0

Added

• Provider model capabilities can be configured per Agent and model — Advanced Provider settings now expose separate Claude Code, Codex, and OpenCode panels for context limits, reasoning behavior, tool and modality support, and protocol-aware OpenCode variants; BigModel's Coding Plan preset uses its documented OpenAI Responses /api/v1 endpoint and glm-5.3 model.
• New sessions follow a global permission-mode default — General settings now choose whether new sessions start in Auto, Read only, or Ask before actions mode; the app defaults to Auto while preserving per-session overrides.

Changed

• OpenCode upgraded to 1.18.23 — The SDK and every bundled platform binary move together from 1.18.4 to 1.18.23 as one pinned compatibility unit.

Fixed

• Saved Provider Auth Tokens can be inspected again — Provider editors now show a masked saved-token state and reveal the credential only after an explicit eye-button action through the desktop-authenticated, non-cacheable API path.

Signing status

• macOS: UNSIGNED — full code-signing and notarization credentials were unavailable
• Windows: UNSIGNED — full Azure Trusted Signing credentials were unavailable
• Linux: No platform signing requirement; Linux release assets are published without a platform signature.

Comate 0.3.1

Comate 0.3.1 polishes Agent settings, aligns the desktop and website with the Comate brand, and expands the website's end-to-end Agent workflow guidance.

Added

• Complete Agent workflows on the Comate website — The bilingual product site now walks through a controlled finance-report task, presents current desktop product evidence, offers platform-specific downloads, and measures visits and download actions only after consent.

Changed

• Clearer, more compact Agent settings — Claude Code options now live in a polished collapsible group with responsive spacing. Output style is an app-global preference that applies consistently across sessions.
• Shared Comate brand colors — The desktop and website now use a shared logo-derived color system for primary actions, activity, attention states, work surfaces, and supporting visuals.

Fixed

• First-click sidebar collapse — The left sidebar now responds immediately to its collapse control.
• Stronger website release and analytics checks — Download trust decisions, consent handling, release metadata, and shared bilingual content now have stricter automated coverage.

Signing status

• macOS: UNSIGNED — full code-signing and notarization credentials were unavailable
• Windows: UNSIGNED — full Azure Trusted Signing credentials were unavailable
• Linux: No platform signing requirement; Linux release assets are published without a platform signature.

Comate 0.3.0

Comate 0.3.0

This release completes the Electron desktop transition and expands the day-to-day session workflow.

Highlights

• Choose Claude Code output styles directly from the prompt toolbar and see structured context usage plus reasoning effort in the status bar.
• Send screenshots and images in prompts, and preview workspace video and audio files inside Comate.
• View live coding-plan usage for supported providers from the account menu.
• Use workspace context menus, activity-based workspace ordering, atomic prompt reference chips, and smoother file/browser panels.
• Run the bundled Claude Code 2.1.237 toolchain with its reliability and security fixes.

Desktop release reliability

• Automatic update checks now surface feed failures instead of reporting false success.
• Electron updater manifests and blockmaps are generated and validated even when macOS or Windows signing credentials are unavailable.
• Unsigned macOS and Windows installers remain available as direct assets; the managed signing-status section below records the exact status of this build.
• The release workflow validates packaged renderer files, updater feed configuration, draft assets, and version consistency before publication.

See the full changelog for all additions, changes, and fixes.

Signing status

• macOS: UNSIGNED — full code-signing and notarization credentials were unavailable
• Windows: UNSIGNED — full Azure Trusted Signing credentials were unavailable
• Linux: No platform signing requirement; Linux release assets are published without a platform signature.

Comate v0.2.0

Comate 0.2.0

Comate 0.2.0 builds on the Electron desktop line with a fully integrated New Chat flow, detachable embedded-browser windows, and a large set of reliability fixes for packaged builds and enterprise bots.

Warning

The macOS and Windows installers in this release are unsigned because signing credentials are unavailable. macOS Gatekeeper and Windows SmartScreen may show warnings. No macOS/Windows auto-update manifests are included; update those installations manually.

Highlights

• New Chat now starts conversations directly from a workspace: choose an existing workspace or create one from the composer header, use the same Skills, Files, and Provider controls as an active chat, and submit the first prompt to create and open the session immediately.
• Embedded browsers can open in an independent desktop window pinned to their originating chat, while the main panel keeps focus and restore controls.
• Enterprise chat bots (WeCom, Feishu) explain progress with grouped, user-facing activity states instead of raw tool names, keep approval and question context in read-only cards, and cancel pending decisions safely.
• Browser page understanding and element discovery are now explicit: a text-only, token-bounded semantic outline with getPageState → findElements → getElementDetails → act; the old inspectElement is renamed getElementDetails, and snapshot becomes the visual-only takeScreenshot.
• Edit tool changes now render as one unified diff instead of separate before/after blocks.
• Process Region drawer: smooth open/close and expand/collapse motion, resizing up to 800px, chat font-size support, and tool headers aligned with reasoning rows; subagent lists match the same presentation.

Reliability fixes

• New Chat no longer crashes the packaged backend while naming a session (the entire Intl.Segmenter dependency chain was purged from the packaged runtime; the sidecar build now smoke-tests session creation with English, Chinese, and truncating prompts).
• Scheduled-task creation validates the schedule before persisting, so invalid input no longer leaves orphaned tasks.
• Bot and scheduled agent sessions no longer inherit ambient Comate credentials from a parent process.
• The file browser previews images (PNG, JPEG, GIF, WebP, AVIF, BMP, ICO, SVG).
• Multi-question prompts submit successfully instead of failing with "Approval is no longer pending."
• Embedded browsers identify and edit ProseMirror-style rich-text editors reliably.
• Prompt send controls stay inside the composer at narrow widths, and the New Chat composer no longer doubles its inner padding.
• Interrupted turns no longer show internal EDE diagnostics; interrupted sub-agent timers stay fixed in inactive sessions.
• Comate exits duplicate desktop launches immediately, restoring the existing window.
• Night-idle todos explain missing workspace setup and reactivate when a workspace is selected.

Comate v0.1.0

Comate 0.1.0

This is Comate's first Electron-based desktop release. It preserves the existing React UI, Express sidecar, and application data directory while replacing the Tauri shell and browser runtime.

Warning

The macOS and Windows installers in this release are unsigned because signing credentials were unavailable. macOS Gatekeeper and Windows SmartScreen may show warnings. No macOS/Windows auto-update manifests are included; update those installations manually.

Highlights

• Migrated the desktop shell from Tauri to Electron, with native macOS, Windows, and Linux packaging.
• Replaced the bundled Chrome for Testing / Steel stack with native, per-session isolated browser views, reducing the application footprint by roughly 267 MB.
• Added Linux distribution through AppImage and Debian packages. AppImage supports in-app updates; .deb updates require downloading and reinstalling the package.
• Made Claude the default agent for new installations.
• Collapsed tool cards and thinking blocks by default, with bounded expanded content and search-aware expansion.
• Added Electron updater and migration infrastructure for existing installations.

Reliability fixes

• Prevented duplicate desktop instances and restored the existing window when Comate is launched again.
• Fixed packaged releases opening to a blank window by compiling and validating the renderer bundle during release builds.
• Fixed Windows CDP cleanup failures caused by Chromium files remaining locked during process shutdown.
• Waits for the sidecar API to become ready before loading the desktop UI.
• Added the missing top frame to restored Windows windows.
• Fixed browser handoff approval cards remaining pending after the embedded browser closes.
• Fixed todo-launched sessions not appearing and opening immediately.

Upgrade notes

• UI preferences stored in the old webview, such as panel visibility and pane widths, reset once after upgrading.
• Embedded-browser site sessions do not migrate from the previous runtime. Sites saved with Remember this site can restore their login; other sites require a fresh login.
• Legacy browser profiles, pidfiles, and downloaded browser caches are removed on first launch. Remembered-site login data is preserved.
• On Windows, the installer handles migration from the previous Tauri installation.

Downloads

• macOS Apple Silicon: Comate-0.1.0-mac-arm64.dmg
• macOS Intel: Comate-0.1.0-mac-x64.dmg
• Windows x64: Comate-0.1.0-win-x64.exe
• Linux x64: Comate-0.1.0-linux-x86_64.AppImage or Comate-0.1.0-linux-amd64.deb
• The macOS .zip files are updater payloads for their corresponding architectures, not additional installers.

All Linux, Windows, and macOS build jobs passed in the final 0.1.0 release build.

Full changelog: v0.0.33...v0.1.0

Comate v0.0.33

Comate v0.0.33

This release improves the WeCom bot sandbox permission model on degraded hosts, tightens the todo creation shortcut, and makes the degraded-sandbox warning banner dismissible.

🆕 Changed

• Owner/admin-approved Bash on degraded hosts (bot sandbox model, AE5) — On a host where the execution sandbox is unavailable (notably Windows, where the probe reports platform-unsupported), a regular WeCom bot member's Bash command no longer fails outright with "repair sandboxing first." The permission gate now routes it directly to the channel's owner/admin approval cards — the same flow as an out-of-sandbox escape — and runs it unsandboxed once an owner or admin approves. Owner/admin members still bypass approval, non-WeCom channels (e.g. Feishu) still deny until their card flow is aligned, and the existing per-user/per-bot caps, dedupe, and always-allow persistence bound the request volume. The authorization is decided by the gate from host sandbox state, not by the model's per-call sandbox flag.

• Todo title submission shortcut — Creating and renaming todos now requires Ctrl/Cmd+Enter instead of plain Enter.

• Dismissible sandbox warning — The degraded sandbox warning banner can now be dismissed for the current app session without changing the sandbox posture.

📥 Downloads

Platform	Asset
macOS (Apple Silicon)	Comate_0.0.33_aarch64.dmg
macOS (Intel)	Comate_0.0.33_x64.dmg
Windows (x64)	Comate_0.0.33_x64_en-US.msi

Existing installations will pick up this version automatically via the built-in updater.

Full Changelog: v0.0.32...v0.0.33

---

Comate v0.0.33(中文)

本次发布优化了降级主机上的 WeCom 机器人沙箱权限模型,收紧了待办事项的提交快捷键,并让降级沙箱警告横幅可被关闭。

🆕 变更

• 降级主机上的所有者/管理员审批式 Bash(机器人沙箱模型,AE5) —— 当执行沙箱不可用时(尤其是 Windows,探测结果为 platform-unsupported),普通企业微信机器人成员的 Bash 命令不再直接以"请先修复沙箱"报错。权限网关现在会将其直接路由到频道的所有者/管理员审批卡片——与沙箱逃逸走同一条流程——经所有者或管理员批准后即以非沙箱方式执行。所有者/管理员成员仍可绕过审批,非企业微信频道(如飞书)在卡片流程对齐前仍会拒绝。现有的按用户/按机器人上限、去重和"始终允许"持久化仍然约束请求量。授权由网关根据主机沙箱状态决定,而非模型每次调用的沙箱标志。

• 待办标题提交快捷键 —— 创建和重命名待办事项现在需要 Ctrl/Cmd+Enter,而不是直接按 Enter。

• 可关闭的沙箱警告 —— 降级沙箱警告横幅现在可以在当前应用会话内关闭,而不会改变沙箱状态。

📥 下载

平台	文件
macOS(Apple Silicon)	Comate_0.0.33_aarch64.dmg
macOS(Intel)	Comate_0.0.33_x64.dmg
Windows(x64)	Comate_0.0.33_x64_en-US.msi

已安装的版本会通过内置自动更新器自动获取本次更新。

完整变更日志:v0.0.32...v0.0.33

Comate v0.0.32

Comate v0.0.32

This release adds provider usage to the status bar, switches font-size controls to direct numeric entry, and fixes cross-workspace session reuse, WeCom interactive-card updates, and privileged bot Bash access.

🆕 Added

• Provider usage in the status bar — The status bar now shows provider usage information without leaving the active workspace.

🔁 Changed

• Numeric font-size controls — Chat and UI font sizes can now be entered directly as bounded pixel values instead of choosing from preset sizes.

🐛 Fixed

• Workspace-scoped chat sessions — Starting a chat no longer reuses a session from another workspace.
• WeCom interactive card updates — Approval and workspace cards now carry the required task identity and update using valid interactive-card payloads.
• Privileged bot Bash access — Bot owners and admins now receive their intended Bash approval bypass.

📥 Downloads

Platform	Asset
macOS (Apple Silicon)	Comate_0.0.32_aarch64.dmg
macOS (Intel)	Comate_0.0.32_x64.dmg
Windows (x64)	Comate_0.0.32_x64_en-US.msi

Full Changelog: v0.0.31...v0.0.32

---

Comate v0.0.32(中文)

本次发布在状态栏加入了 Provider 用量展示,字号控件改为直接输入数值,并修复了跨工作区会话复用、企业微信互动卡片更新以及特权机器人的 Bash 访问问题。

🆕 新增

• 状态栏中的 Provider 用量 —— 状态栏现在会显示 Provider 用量信息,无需离开当前工作区。

🔁 变更

• 数值字号控件 —— 聊天与界面字号现在可以直接输入带边界的像素值,而不再从预设档位中选择。

🐛 修复

• 工作区作用域的聊天会话 —— 新建聊天不再复用来自其他工作区的会话。
• 企业微信互动卡片更新 —— 审批卡片与工作区卡片现在携带必需的任务标识,并使用有效的互动卡片载荷进行更新。
• 特权机器人 Bash 访问 —— 机器人所有者与管理员现在能获得原本预期的 Bash 审批绕过。

📥 下载

平台	文件
macOS(Apple Silicon)	Comate_0.0.32_aarch64.dmg
macOS(Intel)	Comate_0.0.32_x64.dmg
Windows(x64)	Comate_0.0.32_x64_en-US.msi

完整变更日志:v0.0.31...v0.0.32

Comate v0.0.31

Comate v0.0.31

A major release: a browser API-discovery workbench, big expansions to Skills (provider filters/health, WeSkillHub, Enterprise Zone, Expert Packages), and a wide-ranging hardening of the WeCom bot sandbox — MCP tool classification, loopback API authentication, decision audit trail, an escalation approval ledger, and remote owner/admin approval cards for regular members.

🆕 Added

• Browser API discovery workbench — Agents can inspect one selected DOM element, bracket one browser action with bounded network capture, and receive ranked credential-redacted API candidates in chat. A task-scoped authenticated-request MCP and new comate api request CLI replay selected APIs through Comate-held cookies or bearer credentials without exposing them; GET/HEAD calls remain fluid, mutations show an exact sanitized approval card, and explicit Remember-site consent can preserve the selected authentication for browser-closed reuse.

• Skill search provider filters, health, and views — Skill Search now lets users include or exclude any connected provider, remembers that choice globally, and identifies unavailable providers with safe failure reasons and per-provider Retry. Searches skip unavailable sources while preserving healthy results, clearly warn when result coverage is incomplete, and support persistent card and list result layouts.

• WeSkillHub federated Skills provider — WeSkillHub is now the fifth federated provider, with discovery using the generic project/global Skill installation and update lifecycle.

• Enterprise Zone in Skills — Skills now includes a top-level Enterprise Zone for searchable, industry-filtered enterprise discovery, per-enterprise Skill search and sorting, Skill details, and individual project/global installation through the standard Skill flow.

• Expert Packages in Skills — Skills now includes a top-level Expert Packages area with searchable and scene-filtered package discovery, package and included-Skill detail pages, and in-app project/global installation. Complete package installs preserve partial successes and support failed-item retry, while package orchestration remains runtime-compatible but is labeled separately from standard Skills.

• MCP tool classification and admin boundary (bot sandbox model, U9) — MCP tools in bot sessions are no longer silently allow-all: every mcp__* call is now classified at the permission gate as read / write / unknown. Server-provided annotations (readOnlyHint / destructiveHint) are honored first, a per-server override in the bot policy (mcpClassification, keyed by server name) wins over the annotation, and an unclassifiable tool always asks — it can never fall through to allow. Write-class and unknown-class calls enter the escalation flow as mcp-write requests: regular members route to the channel's owner/admin approval cards (with the same dedupe, caps, TTL, and no-approver deny-with-explanation as bash escapes), owner/admin requesters self-ask, and Feishu channels keep the previous deny until their card flow is aligned. Read-class tools continue through the existing category policy without a card. The bundled scheduled-tasks MCP server now advertises annotations (list_scheduled_tasks is read-only; task creation is write-class), and the admin role's target boundary is enforced at the gate itself: an admin can write the workspace and the closed capability set (.claude/skills, .claude/agents) but is now denied writes into .claude/plugins, hooks, settings files, and anything outside the workspace — previously only bash was constrained (by the sandbox), while the Edit/Write tools ran wide open inside .claude for admins. The owner's reach is unchanged.

• Loopback API authentication (bot sandbox hardening, U12) — the entire /api surface is now default-deny authenticated at the route-registration layer: every present and future route requires a Bearer credential, with exemptions declared explicitly. Bot sessions receive a per-session capability token (24h TTL, rotated on runtime rebuild, revoked on session close/demotion/boot, stored as SHA-256 hashes) that reaches only the closed set of wecom CLI routes, with identity and admin rights derived from the token-bound session — a self-asserted sessionId is no longer trusted. The desktop client authenticates with a per-boot GUI credential delivered via the sidecar ready message (Tauri) or a 0600 credential file (dev proxy). The wecom CLI context moved from the workspace-root .claude/wecom-context.json (discovered by walking up directories) to a per-session data/<user>/.runtime/wecom-context.json passed explicitly via COMATE_WECOM_CONTEXT_FILE, so a context planted in a user-writable directory can no longer redirect the CLI; the CLI also routes its loopback calls through the sandbox HTTP proxy when one is present (required for egress from sandboxed sessions).

• Bot decision audit trail (bot sandbox hardening, U6) — the bot audit log now records the sandbox permission model's decision points: bash denials (with the structural reason and routing class), out-of-sandbox requests and their resolutions (approved/denied/expired, with the approver as actor and the requester recorded alongside), passlist rule additions (with manual/approval provenance), admin writes into the workspace capability dirs (.claude/skills, .claude/agents), capability-token mint/revoke, and loopback authentication rejections. Long commands, rules, and domains are stored in full with a SHA-256 integrity hash instead of being redacted, while secret-shaped values (API keys, bearer tokens, the 48-hex capability tokens) are masked at any length; audit rows are retained for 90 days and then purged, and the audit surface stays unreachable for bot session tokens — only the desktop credential can read it.

• Bot escalation approval ledger (bot sandbox hardening, U8) — out-of-sandbox approval requests from owner/admin bot sessions are now recorded in a persistent escalation ledger (requester, recipients, rule payload, state, expiry) instead of living only in memory. Approvals are bounded by a 30-minute TTL when the tool call carries no timeout of its own: an unanswered request now settles as a fail-closed denial with an expired audit row and a timeout notice to the requester (previously it could wait indefinitely). A server restart no longer strands approvals either — every still-pending request is expired at boot (never auto-approved), audited, and the requester is notified once the bot's WeCom connection is back. Each pending also carries an audience marker (self for owner/admin requesters, admins otherwise — fail-safe) that the upcoming remote-approval cards will route on. Desktop approvals now resolve through the same provenance writer as card approvals with identical audit shape, the desktop approvals endpoint never spawns a session runtime just to answer an approval (404 when none is live), and the desktop session list shows its existing pending indicator on bot sessions awaiting approval.

• Remote owner/admin approval cards for regular members (bot sandbox model, U11) — a regular channel member's out-of-sandbox bash request no longer hits a blanket denial: it now creates an escalation that delivers an actionable approval card to the channel's owner/admin. Owner and Admin requests retain their default role bypass and run without approval. The regular-member requester receives a read-only notice card instead of a self-approvable one — self-approval was never supervision. Card clicks are authorized against a fresh role check on the ledger row's bot (a demoted admin loses approval power immediately and members of other bots are rejected), and settlement is transactional first-click-wins, so double-clicks and late/replayed clicks are harmless no-ops; the clicker's card flips terminal while everyone else — the requester on any outcome and the non-clicking approvers — receives a terminal notification card on approve, deny, or TTL expiry. "始终允许" (always allow) now persists an exact-match structural rule (the literal command, never a wildcard) into the bot's out-of-sandbox passlist with the approver's identity and approval provenance in the audit trail, so the same command in a future session runs without asking while a same-tool/different-arguments variant still cannot match; SDK suggestions that would widen scope (mode changes, added directories, rule replacement) are dropped and suppress the button entirely. Escalation volume is bounded: parameter-variant retries collapse into a single pending card, per-user hourly and per-bot outstanding caps fail closed with a notice, and a turn that keeps retrying after repeated denials gets an explicit stop instruction. Regular members' allowUnsandboxedCommands is now enabled — the escape hatch finally has a supervised path. Feishu channels keep the previous behavior until their card flow is aligned.

🔁 Changed

• Incomplete Expert Package installation — package validation is now advisory. The app installs the raw package orchestration and every resolvable included Skill, reports unavailable children as per-item failures, and lets users correct the installed files afterward.

🐛 Fixed

• Packaged Comate CLI startup — the bundled comate executable now contains exactly one shebang, so the packaged CLI starts correctly instead of failing on a duplicated interpreter line.

📥 Downloads

Platform	Asset
macOS (Apple Silicon)	Comate_0.0.31_aarch64.dmg
macOS (Intel)	Comate_0.0.31_x64.dmg
Windows (x64)	Comate_0.0.31_x64_en-US.msi

Full Changelog: v0.0.30...v0.0.31

---

Comate v0.0.31(中文)

一次大版本更新:浏览器 API 发现工作台、Skills 的大幅扩展(Provider 筛选/健康、WeSkillHub、企业专区、专家包),以及对企业微信机器人沙箱的全面加固——MCP 工具分类、回环 API 鉴权、决策审计追踪、升级审批账本,以及面向普通成员的远程所有者/管理员审批卡片。

🆕 新增

• 浏览器 API 发现工作台 —— Agent 可以检视单个选中的 DOM 元素,用受限的网络捕获包裹一次浏览器操作...

Comate v0.0.30

[0.0.30] - 2026-07-31

Added

• Inline todo title editing — a todo title in the list can now be renamed in place: single-click still opens the detail pane, double-clicking the title enters edit mode, Enter or Cmd/Ctrl+Enter saves, Escape cancels, and blur saves. Empty titles are rejected (silently reverted). The same 2000-character cap applies.

• Embedded browser: internal HTTP and insecure-certificate sites — the embedded browser now respects explicit http:// URLs instead of automatically upgrading them to HTTPS, preventing upgrade/downgrade redirect loops that surfaced as ERR_BLOCKED_BY_CLIENT. Private-CA and hostname-mismatched sites can also load through a global Settings → General toggle, persisted server-side and applied to every spawned Chrome as --ignore-certificate-errors; it is on by default and can be disabled for strict certificate validation. (Replaces the earlier COMATE_BROWSER_IGNORE_CERT_ERRORS env var, which is now a no-op.) Applies to the embedded browser only.

• GitHub Issues sync — todos are now global, shareable entities that sync with GitHub Issues. Connect a GitHub account (GitHub App via Device Flow, or a fine-grained PAT), then publish a local todo to an issue or pull an issue into a local replica. Sync is origin-anchored and field-class: comments merge append-only both ways, status/labels/assignee accept the remote, and title is origin-wins with both-sides-edited conflicts surfaced for an accept-local/accept-remote choice (never auto-clobbered). Sync runs on-demand when the Todos panel opens or you click refresh; a single-flight guard collapses overlapping triggers. A remote deletion is detected and flagged, never silently destroying local comments. Tokens are encrypted at rest and never appear in any response or log.

Changed

• Panel background color redesign — simplified the main layout to a minimal white-and-gray palette. App header, left sidebar, and status bar now share a single chrome gray surface, while the chat header, message body, right panel, and prompt input share a single work white surface. App header uses a subtle shadow instead of a bottom border to avoid a double divider with the chat header. Added design comparison prototypes under docs/design/.

• Right sidebar browser integration — the embedded browser is now a tab inside the right sidebar alongside Files and Git Changes. The old chat-header browser button and the separate in-chat browser pane are removed. The right sidebar now collapses completely (no 40 px rail) and animates its width when expanding or collapsing. Browser iframes stay mounted across tab switches and collapse/expand so the cast stream never reloads.

• Todos panel top-bar layout — the todos panel is now a top-bar task list: the left view rail is removed, smart views (Inbox / Today / Upcoming / All) live in a segmented header control with counts, search filters the active view, and the list spans the full width with information-rich rows showing due dates, labels, and origin. Group-by moved into the header, quick-add sits at the top of the list, and empty/no-results/load-failure states are split so sync errors are visible.

• Todo detail panel — the detail pane is now wider by default and resizable via a left-edge drag handle. It animates when a todo is selected, shows the workspace name instead of its ID, and lets you reassign the workspace and edit the status directly. Linked sessions are clickable and jump to the session. Search and group-by inputs are sized to 34 px.

Fixed

• Embedded browser content appearing scaled down — live browser frames now render at their native 100% CSS-pixel size instead of being fitted to the panel height. When the page viewport is larger than the embedded browser, the viewer shows scrollbars rather than shrinking the content.
• Claude SDK Sessions becoming idle while background agents still run — Session activity now follows the SDK's complete background-task snapshot instead of ending with the main-agent result. Background work keeps the Session active, visible, retained, and stoppable while leaving the composer available for replies and follow-up prompts; the Session list shows a task count and the open chat shows task details. Stop now fences the whole Session immediately, drains late tasks, and closes the SDK process when cancellation cannot be proven.
• Individual Claude background task controls — each running background agent or command can now be stopped independently from the activity panel without interrupting the foreground turn or other background work.
• First chat message flush against the header — the conversation content area now has top padding, so the first message no longer sits directly on the chat header's bottom border; the 16 px gap matches the existing spacing below each message row.

Comate v0.0.29

Comate v0.0.29

修复 / Fixed

• 企微明文用户 ID 解析：修复了明文用户 ID 无法解析的问题，自动解析与手动"解析待处理"均受影响。统一 schema 迁移已将企业 corp 凭证从 workspace.settings 迁移至机器人企微渠道配置并删除了原字段，而解析器仍在读取被删除的字段，导致解析静默失败。现已改为从渠道配置读取凭证，自动与手动解析均可正常工作。

  WeCom plaintext user ID resolution — resolves correctly again, both via the background auto-flush and the manual "resolve pending" action. The resolver was reading corp credentials from workspace.settings fields that the unified-schema migration deletes; it now reads them from the bot's wecom channel config (their current home), so resolution no longer fails with "credentials not configured".

下载 / Downloads

• macOS (Apple Silicon): Comate_0.0.29_aarch64.dmg
• macOS (Intel): Comate_0.0.29_x64.dmg
• Windows: Comate_0.0.29_x64_en-US.msi

---

完整变更日志 / Full changelog: https://github.com/ai-dvps/comate/blob/main/CHANGELOG.md