qwen-code/CHANGELOG.md

Changelog

All notable changes to Qwen Code are documented in this file.

The format is based on Keep a Changelog, and the project follows Semantic Versioning. Only stable releases are listed; nightly and preview pre-releases are intentionally omitted.

This file is generated automatically from GitHub Releases. Do not edit it by hand — run npm run changelog to regenerate.

0.26.9 - 2026-11-06

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Skill installation and activation

Skill ZIP uploads, daemon APIs, TypeScript SDK support, bulk activation, and recovery protections make installing and managing Skills more reliable.

• Adds PUT /skills/activation and setUserSkillsActivation to enable or disable 1-100 user Skills in one request, with saved preferences and propagation status. (#16186)
• Prepares Skill installation for larger archives by adding disk-based ZIP handling and recovery-backup protection to the installer. (#16421)
• Adds daemon API routes for installing large Skill ZIPs by upload or host file path and querying installation receipts after failures. (#16422)
• Adds TypeScript SDK support for installing Skill ZIPs from File, Blob, bytes, or host path, with attempt IDs and receipt lookup. (#16423)
• WebShell can now upload Skill ZIP archives from the Skills page, show daemon limits, and check uncertain installation results. (#16424)

Web Shell sessions and composer

Web Shell session titles, Add menu polish, draft preservation, batch deletion, and listSessions pagination fixes improve everyday session management.

• Keeps Web Shell sidebar session titles visible when a transcript is rebuilt, preventing session names from disappearing. (#16299)
• Polishes the Web Shell composer Add menu with consistent icons, labels, descriptions, spacing, and improved accessibility for actions and Plan state. (#16181)
• WebShell now preserves unsent drafts when ArrowUp or ArrowDown moves through wrapped prompt lines before history navigation. (#14909)
• web-shell: batch session deletions at the route limit (#14637)
• listSessions pagination no longer skips sessions when multiple session files share the same modification time. (#16024)

Managed agents and sessions

Managed agent automation, durable questions, activation renewals, writer recovery, and history caching improve hosted session reliability and speed.

• managed-agent: support durable structured user questions (#16433)
• managed-agent: enable automation by default (#16448)
• Managed agent activation renewals now locate the activation lease more efficiently, reducing renewal lookup overhead. (#16336)
• Improves managed-agent performance by removing tenant guard locks from native journal snapshot publication. (#16410)
• Managed Session history rebuilds now cache message bodies per process, reducing repeated reads and improving performance for hosted turns. (#16337)
• HTTP Managed Session writer recovery now keeps a live in-flight renewal grant instead of unnecessarily re-acquiring the writer. (#16287)

GitHub PR automation

Sandboxed /verify can read PR review threads, /autofix-pr refusals expose machine-readable codes, and GitHub inbound recovery retries failed reads.

• The sandboxed /verify round can now see PR review threads, including resolved and outdated threads, so reviewer findings are available during verification. (#16300)
• /autofix-pr helper refusals now include a machine-readable refused code and next step on stderr, making failures easier to handle. (#16238)
• GitHub inbound recovery now retries after a failed read instead of stopping, so recovered tasks run once the task store becomes readable. (#16402)

Model and provider support

Image input fixes for Llama, LLaVA, Pixtral, Mistral Small, and Qwen VL models plus OCI Generative AI documentation broaden provider coverage.

• Enables image input for Llama 3.2 Vision, Llama 4, LLaVA, Pixtral, and Mistral Small 3.1 and 3.2 models. (#16311)
• core: accept images for Qwen VL models under their Ollama names (#16309)
• providers: document OCI Generative AI under OpenAI-compatible providers (#16345)

Channels and messaging

QQ proactive delivery and richer Channel event timeline recording improve active group messaging, delivery tracking, and scan health visibility.

• Enables QQ proactive delivery so loop messages, webhook tasks, background responses, and daemon sends can be sent as active group messages. (#15635)
• Channel event timeline now records private bound-session deliveries, undelivered events, and scan health outcomes. (#16237)

CLI and terminal status

Interactive CLI reports working, blocked, done, or error status via OSC 7501, while experimental.unifiedAgentLoop helps Ink TUI finish tool batches.

• The interactive CLI now reports working, blocked, done, or error status to terminals using OSC 7501 with app=qwen-code. (#16407)
• With experimental.unifiedAgentLoop enabled, Ink TUI turns await their shared tool batch so tool continuations complete within the turn. (#16385)

Daemon, runtime, and account recovery

Faster SSH registration and MCP startup, reduced tool-execution locking, bounded daemon replay memory, and subscription recovery improve platform stability.

• SSH registration settlement and Session worktree preparation are less serialized, allowing more concurrent Runtime work and faster workspace preparation. (#16386)
• Loads MCP client Google auth providers on demand to reduce startup load while keeping configured server behavior unchanged. (#16242)
• Improves ordinary session tool execution performance by reducing unnecessary locking during admission and dispatch. (#16415)
• Daemon command and skill replay now limits snapshot memory with an 8 MiB per-session ring, configurable via --event-ring-max-bytes. (#16413)
• subscription: isolate activation and restore sign-in recovery (#16453)

Other Changes

• Internal test coverage for renewal recovery and writer seal handling was hardened without changing production behavior. (#14482)
• Consolidates Web Shell sidebar test scaffolding onto a shared sidebarHarness without changing user-visible behavior. (#16016)
• Adds a design proposal for moving PR state transition detection from GitHub channel polling to the daemon PR refresh timer. (#16246)
• Shortens the stale PR policy to mark pull requests stale after 30 inactive days and close them after 14 more days, with a larger sweep budget. (#16435)
• Fixes release Docker integration builds to use the same apt-mirror probe as E2E builds, improving mirror selection and cache reliability. (#16278)
• The no-AK integration gate now runs on the self-hosted ECS pool for every pull request, including untrusted forks. (#16132)
• test(managed-agent): cover pending sibling approval holds (#15994)
• ci: build the PR tree once and shorten the required path (#16450)

中文摘要

Skill 安装与激活

Skill ZIP 上传、daemon API、TypeScript SDK 支持、批量激活与恢复备份保护，让 Skill 安装和管理更可靠。

• 新增 PUT /skills/activation 与 setUserSkillsActivation，可一次启用或禁用 1-100 个用户 Skill，并返回保存与传播状态。 (#16186)
• 为大型 Skill 压缩包安装做准备，在安装器中加入从磁盘读取 ZIP 的能力与恢复备份保护。 (#16421)
• 新增守护进程 API，可通过上传或主机文件路径安装大型 Skill ZIP，并在失败后查询安装回执。 (#16422)
• TypeScript SDK 支持通过 File、Blob、字节或主机路径安装 Skill ZIP，并提供尝试 ID 与回执查询。 (#16423)
• WebShell 现在可从 Skills 页面上传 Skill ZIP，显示守护进程限制，并查询不确定的安装结果。 (#16424)

Web Shell 会话与输入

Web Shell 会话标题、Add 菜单细节、草稿保留、批量删除与 listSessions 分页修复改进日常会话管理。

• 当 transcript 重建时保留 Web Shell 会话列表标题，避免会话名称丢失。 (#16299)
• 优化 Web Shell 作曲器 Add 菜单的图标、标签、描述、间距，并改进无障碍展示与 Plan 状态指示。 (#16181)
• WebShell 现在在 ArrowUp 或 ArrowDown 浏览换行提示行时保留未发送草稿。 (#14909)
• web-shell: batch session deletions at the route limit (#14637)
• listSessions 分页在多个会话文件修改时间相同时不再跳过会话。 (#16024)

Managed agent 与会话

Managed agent 自动化、持久用户问题、activation 续期、writer 恢复与历史缓存提升托管会话的可靠性与性能。

• managed-agent: support durable structured user questions (#16433)
• managed-agent: enable automation by default (#16448)
• Managed agent 激活续期现在能更高效定位 activation lease，降低查询开销。 (#16336)
• 通过移除 native journal 快照发布中的 tenant 守卫锁，提升 managed-agent 性能。 (#16410)
• Managed Session 历史重建现在按进程缓存消息正文，减少重复读取并提升性能。 (#16337)
• HTTP Managed Session 写入方恢复现在会保留有效的在途续期授权，避免不必要重新获取。 (#16287)

GitHub PR 自动化

沙箱 /verify 可读取 PR review threads，/autofix-pr 拒绝输出机器可读代码，GitHub 入站恢复在读取失败后重试。

• 沙箱 /verify 轮次现在可读取 PR review threads，包括已解决和过期讨论，便于验证时参考评审意见。 (#16300)
• /autofix-pr 助手拒绝时会在 stderr 输出 refused 代码和 next 步骤，便于自动处理。 (#16238)
• GitHub 入站恢复在读取失败后会继续重试，任务存储可读后已恢复任务将运行。 (#16402)

模型与 Provider 支持

Llama、LLaVA、Pixtral、Mistral Small 与 Qwen VL 图像输入修复，以及 OCI Generative AI 文档扩展 provider 覆盖。

• Llama 3.2 Vision、Llama 4、LLaVA、Pixtral 及 Mistral Small 3.1/3.2 支持图像输入。 (#16311)
• core: accept images for Qwen VL models under their Ollama names (#16309)
• providers: document OCI Generative AI under OpenAI-compatible providers (#16345)

Channel 与消息

QQ 主动消息投递与更完整的 Channel 事件时间线记录改进群消息、投递跟踪和扫描健康可见性。

• QQ 支持主动消息投递，循环、webhook、后台任务和 daemon 发送可作为主动群消息发出。 (#15635)
• Channel 事件时间线现在记录私有绑定会话投递、未投递事件和扫描健康结果。 (#16237)

CLI 与终端状态

交互式 CLI 通过 OSC 7501 报告 working、blocked、done 或 error 状态，experimental.unifiedAgentLoop 帮助 Ink TUI 完成工具批处理。

• 交互式 CLI 现在通过 OSC 7501 向终端报告 working、blocked、done 或 error 状态，并标记 app=qwen-code。 (#16407)
• 启用 experimental.unifiedAgentLoop 后，Ink TUI 回合会等待共享工具批处理完成。 (#16385)

Daemon、Runtime 与账户恢复

SSH 注册、MCP 启动、工具执行锁与 daemon 回放内存限制优化，以及 subscription 恢复改进平台稳定性。

• SSH 注册结算与 Session worktree 准备减少串行限制，支持更多并发 Runtime 工作。 (#16386)
• 按需加载 MCP 客户端的 Google auth providers，降低启动负载且配置服务器行为不变。 (#16242)
• 通过减少普通会话工具执行准入与调度中的不必要锁，提高执行性能。 (#16415)
• 守护进程命令和技能回放限制快照内存，默认每会话 8 MiB，可用 --event-ring-max-bytes 配置。 (#16413)
• subscription: isolate activation and restore sign-in recovery (#16453)

其他变更

• 增强续期恢复与 writer seal 的测试覆盖，不改变生产行为。 (#14482)
• 将 Web Shell 侧边栏测试脚手架合并到共享 sidebarHarness，不影响用户可见行为。 (#16016)
• 新增设计提案，将 PR 状态变化检测从 GitHub 频道轮询迁移到 daemon 的 PR 刷新定时器。 (#16246)
• 将 stale PR 策略缩短为 30 天无活动标记、再过 14 天关闭，并提高单次清理预算。 (#16435)
• 修复 release Docker 集成构建，使其与 E2E 一样使用 apt-mirror 探测，提升镜像选择与缓存可靠性。 (#16278)
• no-AK 集成门禁现在对所有 pull request 路由到自托管 ECS 池，包括不受信任分支。 (#16132)
• test(managed-agent): cover pending sibling approval holds (#15994)
• ci: build the PR tree once and shorten the required path (#16450)

Complete Change List (39 pull requests)

Features

• verify-pr: give the sandboxed round the PR's review threads (#16300) by @wenshao
• autofix-pr: give helper refusals a code and a next step (#16238) by @wenshao
• qqbot: deliver proactive loop messages as active group messages (#15635) by @Eric-GoodBoy-Tech
• cli: have an Ink TUI turn await its shared tool batch (#16385) by @qqqys
• serve: add atomic user Skill activation (#16186) by @callmeYe
• skills: add file-backed archive installation kernel (1/4) (#16421) by @callmeYe
• skills: expose native archive routes and receipts (2/4) (#16422) by @callmeYe
• sdk: support native Skill archives and receipt lookup (3/4) (#16423) by @callmeYe
• web-shell: polish composer action menu presentation (#16181) by @ytahdn
• managed-agent: support durable structured user questions (#16433) by @doudouOUC
• cli: report interactive program status with OSC 7501 (#16407) by @sorrycc
• web-shell: upload Skill archives and reconcile results (4/4) (#16424) by @callmeYe
• channels: journal private bound-session delivery and scan health (#16237) by @wenshao
• managed-agent: enable automation by default (#16448) by @doudouOUC

Bug Fixes

• channels: keep GitHub inbound recovery armed after a failed read (#16402) by @wenshao
• core: keep a writer grant an in-flight renewal refreshed (#16287) by @wenshao
• core: keep session list titles when a transcript is rebuilt (#16299) by @wenshao
• core: accept images for Llama 3.2 Vision, Llama 4, LLaVA, Pixtral and Mistral Small 3.1 and 3.2 (#16311) by @Shizoqua
• ci: apply the apt-mirror probe to the release docker lane too (#16278) by @yiliang114
• core: make persisted session pagination safe for equal mtimes (#16024) by @yiliang114
• serve: bound command snapshot replay memory (#16413) by @doudouOUC
• web-shell: preserve wrapped drafts during arrow navigation (#14909) by @jianyx1
• web-shell: batch session deletions at the route limit (#14637) by @jianyx1
• core: accept images for Qwen VL models under their Ollama names (#16309) by @Shizoqua
• subscription: isolate activation and restore sign-in recovery (#16453) by @liuxuely0626

Performance

• core: read each Managed message body once per process (#16337) by @wenshao
• managed-agent: index the activation lease lookup (#16336) by @wenshao
• managed-agent: narrow SSH settlement and worktree locking (#16386) by @doudouOUC
• core: load the MCP client's Google auth providers on demand (#16242) by @wenshao
• managed-agent: take native journal snapshots off the tenant guards (#16410) by @doudouOUC
• managed-agent: take tool execution control off the tenant guards (#16415) by @doudouOUC

Documentation

• design: propose PR state transitions from the daemon PR refresh (#16246) by @wenshao
• providers: document OCI Generative AI under OpenAI-compatible providers (#16345) by @fede-kamel

Internal Changes

• test(core): harden renewal waits and writer seal coverage (#14482) by @xxwtiancai
• test(web-shell): consolidate sidebar suite fixtures on the shared sidebarHarness (#16016) by @4ekuct25
• chore(ci): shorten stale PR window to 30 + 14 days (#16435) by @wenshao
• ci: route the no-AK integration gate to ECS for every pull request (#16132) by @yiliang114
• test(managed-agent): cover pending sibling approval holds (#15994) by @jianyx1
• ci: build the PR tree once and shorten the required path (#16450) by @sorrycc

New Contributors

• @xxwtiancai made their first contribution in #14482
• @fede-kamel made their first contribution in #16345

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.8...v0.26.9

0.26.8 - 2026-11-05

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• feat(mods): run prompt.submit hooks in the OpenTUI terminal (#16195) by @qqqys
• feat(cli): run the Ink TUI's model rounds on the shared loop's round (#16206) by @qqqys
• feat(web-shell): collapse channel editor tuning behind advanced settings (#16239) by @wenshao
• feat(web-shell): link GitHub PR/Issue targets in channel events (#16226) by @wenshao
• feat(channels): role-aware action templates for GitHub bound-session delivery (#16258) by @wenshao
• feat(cli): run the Ink TUI's tool batches on the shared loop's batch (#16304) by @qqqys
• feat(mods): let a Mod submit a prompt of its own in headless runs (#16302) by @qqqys
• feat(managed-agent): allow concurrent workspace rewinds (#16369) by @doudouOUC
• feat(channels): edit bound-session action templates on demand (#16349) by @wenshao
• feat(serve): allow configuring the Conversations working-directory root (#16105) by @jkxiao
• feat(cli): end the Ink TUI's shared tool batch on the TUI's own stops (#16365) by @qqqys
• feat(web-shell): link channel event rows to their sessions (#16375) by @wenshao
• feat(mods): run the prompts a Mod submits in the Ink terminal (#16373) by @qqqys
• feat: add Qwen subscription support across CLI, VS Code, and Desktop (#16378) by @liuxuely0626

Bug Fixes

• Makes linter archive downloads resume after stalls so CI lint installs fail less often on slow or throttled releases. (#16067) by @yiliang114
• Repins the RHEL 10 rpm build root to a current CentOS stream10 digest so CI containers start instead of failing on manifest lookup. (#16261) by @yiliang114
• Waits for the RocketMQ port listener to clear after stop, reducing flaky script test failures on shared CI hosts. (#16252) by @qwen-code-dev-bot
• fix(core): preserve Gemini thought signatures for tool replay (#16270) by @doudouOUC
• fix(core): recognize Hugging Face TGI context overflow errors so reactive compaction runs (#16245) by @Shizoqua
• fix(core): accept images for Qwen2-VL, Qwen2.5-VL and QVQ models the catalog does not list (#16247) by @Shizoqua
• fix(daemon): automatically discover config Skills changes (#16262) by @callmeYe
• fix(mods): skip hooks on ui.backdrop instead of failing (#16289) by @qqqys
• fix(cli): derive context tool details from declared schemas (#16292) by @callmeYe
• fix(managed-agent): enforce strict SQL mode per embedded connection (#16288) by @doudouOUC
• fix(ci): widen the built-tree download window for eight shards (#16285) by @yiliang114
• fix(ci): fail Java lanes before disk exhaustion (#16068) by @yiliang114
• fix(managed-agent): use current state for approval changes (#16296) by @doudouOUC
• fix(managed): Reap Git descendants before terminating parents (#16298) by @doudouOUC
• fix(cli): gate project hooks behind an approval dialog (#14748) by @TianYuan1024
• fix(managed-agent): reject dangling Hosted cleanup references (#16275) by @doudouOUC
• fix(web-shell): express the PR watch events dependency in the channel editor (#16236) by @wenshao
• fix(serve): read PR sidecars directly for the bindings snapshot (#16241) by @wenshao
• fix(managed-agent): restore JournalCommit under its class-size ratchet (#16323) by @wenshao
• fix(ci): pin COREPACK_HOME for Linux test shards on cold-HOME runs (#16328) by @yiliang114
• fix(managed-agent): bound runtime thread pools (#16330) by @doudouOUC
• fix(managed-agent): refuse every URL fetch refuses and say why a health probe failed (#16110) by @wenshao
• fix(cli): keep project hook approval env inside the workspace (#16316) by @yiliang114
• fix(verify-pr): post flaky demotions on closed PRs and pin the blank class (#16135) by @wenshao
• fix(managed-agent): find the Flyway scripts in the module jar the server ships them in (#16175) by @wenshao
• fix(managed-agent): harden capacity recovery and clean up Java conventions (#16318) by @doudouOUC
• fix(ci): clear the pinned corepack cache before every install (#16351) by @wenshao
• fix(core): headless project-hook trust requires the folder-trust feature (#16264) by @yiliang114
• fix(web-shell): align welcome models and make UI language sessionless (#16297) by @liuxuely0626
• fix(channels): keep GitHub notifications unread on transient thread errors (#16333) by @wenshao
• fix(cli): stop Agent Host disconnects being undone and report unsaved joins (#16156) by @wenshao
• fix(core): tell the model the date once, from the history (#16362) by @qqqys
• fix(cli): preserve raw context usage estimates (#16360) by @callmeYe
• fix(serve): serve qualified hooks status for untrusted workspaces (#16388) by @yiliang114
• fix(cli): land a Web Shell update on the current latest release (#16374) by @wenshao
• fix(channels): keep GitHub replies hit by a secondary rate limit (#16334) by @wenshao
• fix(cli): forward the folder-trust store into the container sandbox (#16390) by @yiliang114

Performance

• Improves managed-agent performance and stability by queueing ordinary journal commits on the Session row instead of a tenant-wide lock. (#16256) by @doudouOUC
• Narrows managed-agent lock scopes for output, lifecycle, and storage updates to reduce contention and improve throughput. (#16268) by @doudouOUC
• perf(managed-agent): separate workspace reads from durable admission (#16283) by @doudouOUC
• perf(managed-agent): avoid storage locks in read preflight (#16312) by @doudouOUC
• perf(managed-agent): index the Runtime slot lookup by isolation key (#16325) by @wenshao
• perf(managed-agent): share the storage fence in session command admissions (#16329) by @doudouOUC
• perf(managed-agent): take ordinary tool publication and read leases off the tenant lock (#16335) by @doudouOUC
• perf(core): defer workflow tool schemas by default (#16310) by @callmeYe
• perf(managed-agent): narrow queue dispatch and workspace session creation locks (#16327) by @doudouOUC
• perf(managed-agent): take ordinary publication lanes off the placement guard (#16354) by @doudouOUC
• perf(memory): move the main session's auto-memory guidance into a bundled skill (#16303) by @qqqys
• perf(managed-agent): remove ordinary journal and publication storage locks (#16370) by @doudouOUC
• perf(serve): keep the ACP spawn channel out of Managed Runtime workers (#16235) by @wenshao

Documentation

• docs: configure Qwen3.x served under a custom model name (#16342) by @spped2000

Internal Changes

• Keeps ACP parity test runtime state outside the workspace so integration tests compare only intended files and run more reliably. (#16254) by @qwen-code-dev-bot
• Makes the ACP parity workspace guard compare the home directory against its pre-run listing, improving test reliability. (#16266) by @wenshao
• Speeds up CI by sharding Linux unit tests and web-shell smoke tests and reusing cached builds instead of rebuilding each workspace. (#16133) by @sorrycc
• test(web-shell): wait for the restored card before clicking continue (#16286) by @yiliang114
• test(interactive): share one CI-scaled wait policy across the TUI suites (#16276) by @yiliang114
• test(e2e): accept both plan-mode wordings in the ACP case (#16273) by @yiliang114
• test(managed-agent): ratchet persistence and runtime-broker class sizes (#16284) by @wenshao
• test(acp): expect the scheduler's plan-mode notice on the shared loop (#16305) by @qqqys
• test(sdk): read before an approved write in the tool-control E2E tests (#16366) by @qqqys
• test(managed-agent): declare v2 reader in the cleanup-conflict MySql IT (#16372) by @yiliang114
• test(managed-agent): fix busy-storage assertion broken by #16369 (#16382) by @yiliang114
• test(ci): pin the zero-skip timeout shape in awaitTwoSkipLines (#16384) by @yiliang114
• test(web-shell): pin the rewind capacity-latch re-open (#16358) by @yiliang114
• test(managed-agent): gate lifecycle-v2 close with a stopped worker (#16123) by @wenshao
• test(review): pin the carried-ledger carrier order and fail-open arms (#10464) (#16395) by @yiliang114

New Contributors

• @sorrycc made their first contribution in #16133
• @spped2000 made their first contribution in #16342
• @liuxuely0626 made their first contribution in #16297

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.7...v0.26.8

0.26.7 - 2026-11-04

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• Adds an operator-scoped model.catalogRefresh setting and lets privacy.usageStatisticsEnabled: false stop background catalog refreshes while local metadata remains available. (#16066) by @yiliang114
• feat(managed-agent): support full model configuration at session creation (#16153) by @doudouOUC
• When experimental.unifiedAgentLoop is enabled, ACP Goal turns run on the shared agent loop with shared rounds, tool batches, and stop points. (#16131) by @qqqys
• Adds a Channel events page to Web Shell with a filterable, auto-refreshing timeline of Channel activity, skip reasons, replies, task transitions, and poll results. (#16106) by @wenshao
• Adds a local /verify-pr Docker recipe that runs untrusted PR code in a container while keeping model credentials on the host through a chat-completion proxy. (#16077) by @wenshao
• Runs Mods prompt.submit hooks in the Ink terminal before typed prompts are processed, allowing hooks to see composer text with pasted content expanded. (#16180) by @qqqys
• Adds a managed Web Shell host entry, improves managed empty-state layout, and consolidates session-failure and unconfirmed-submission notices above the composer. (#15928) by @yiliang114
• Adds Settings → Connections → Allow remote connections to qwen serve, enabling remote access, token use, and client disconnects without restarting. (#15887) by @yiliang114
• With experimental.unifiedAgentLoop enabled, ACP cron and background-notification turns now use the shared agent loop for consistent rounds, tool batching, and stopping. (#16190) by @qqqys
• Workspace uploads can now proceed concurrently while still rejecting occupied destinations and preserving directory occupancy protections. (#16212) by @doudouOUC
• GitHub events for bound PRs now go to the associated session as untrusted prompts, and replies stay private unless published. (#16117) by @wenshao
• Web Shell channel editor can now edit GitHub repository access, mention behavior, and notification polling settings instead of requiring manual settings changes. (#16149) by @wenshao

Bug Fixes

• Pins the RHEL 10 EPEL bootstrap RPM and makes desktop smoke teardown treat Linux zombie processes as stopped while still reporting live daemon failures. (#16154) by @yiliang114
• Compacts live ACP tool-result displays for session replay, so large file edit results show a saved-session preview instead of full old and new contents. (#16172) by @wenshao
• Fixes PR review endpoint routing by selecting the cohort before the model and reports the final bucket, cohort, model, and endpoint hostname. (#16151) by @yiliang114
• fix(serve): keep record_artifact guidance out of Managed Runtime worker results (#16148) by @wenshao
• Binds web-shell config synchronization to the current attachment so config writes made by the first prompt are no longer silently discarded. (#16170) by @wenshao
• Fixes tool-choice failures on Claude Opus 5.5, Sonnet 5.5, Fable 5.1, and Mythos 5.1 by avoiding unsupported forced tool selection. (#15558) by @Adkid-Zephyr
• Fixes Web Shell pasting so a single copied image with an HTTP(S) URL attaches the image, while unavailable file pastes show a warning. (#16179) by @ytahdn
• Requires explicit folder trust before repository-supplied hooks, permission grants, and project language servers can run, improving safety for untrusted projects. (#16182) by @yiliang114
• Anthropic-compatible requests now honor the configured transport timeout and dispatcher, preventing premature request aborts when a longer API timeout is set. (#16193) by @yiliang114
• Remote-access error handling now hides internal error details from callers and fixes several reviewed reliability and security defects. (#16199) by @yiliang114
• On touch devices, navigation now keeps focus on the chat pane instead of opening the composer keyboard, while desktop focus behavior remains unchanged. (#16198) by @ytahdn
• The bundled Generated UI MCP server now runs correctly in Node mode when the CLI is launched from an Electron-based IDE companion. (#16064) by @yar3333
• ACP bridge operations that finish alongside long-lived session or channel events now clean up their results instead of retaining them. (#15347) by @Sumire-no-kai
• Ordinary workspace execution, member release, and file rewind no longer take the tenant-wide placement lock, reducing cross-workspace blocking. (#16205) by @doudouOUC
• Embedded-webview bundle guard failures now show a readable assertion naming the forbidden marker instead of crashing on a large diff. (#16210) by @yiliang114
• Project hooks in .qwen/settings.json now require explicit folder or workspace trust in interactive sessions and show a warning when withheld. (#16196) by @yiliang114
• OpenRouter IDs with :exacto now keep the base model name, so qwen/qwen3-coder:exacto gets its 262,144-token window instead of 200K. (#16165) by @Shizoqua
• Shell permission analysis now treats grep -n and grep --line-number as flags, so grep file reads are still checked. (#14383) by @kabishou11
• Monitor and workflow completion notifications are attributed to the registering turn, letting active-turn results reach the next model request sooner. (#16177) by @yiliang114
• Makes CI sandbox image builds more reliable on hosts with slow default Debian mirrors by switching to a faster mirror only when speed checks require it. (#16218) by @yiliang114
• Scopes GitHub, GitLab, and DWS polling cursors to each workspace, preserving existing positions on upgrade while storing channel state with stricter permissions. (#16231) by @wenshao
• Improves CI apt mirror selection by preferring the VPC intranet mirror and switching when a candidate is fast enough, even if the default mirror seems acceptable. (#16260) by @yiliang114
• Fixes RHEL 10 desktop smoke tests by running them headless without requiring Xvfb. (#16211) by @yiliang114

Performance

• Managed Agent resident Runtime Session claim and release no longer wait on the tenant placement lock, improving resident runtime performance without changing success behavior. (#16217) by @doudouOUC
• Managed Agent turn submission admission and migration preparation avoid the tenant placement lock where safe, reducing wait time without changing request outcomes. (#16223) by @doudouOUC
• Standalone Harness admission, recovery, cleanup, and Hook admission checks no longer take the tenant placement lock, improving Managed Agent performance with unchanged verdicts. (#16221) by @doudouOUC
• Speeds up local Workspace file reads, uploads, and status checks by reusing CLI capability probes when local CLI identity is unchanged. (#16229) by @doudouOUC
• Improves Session responsiveness by reducing locking for common commands such as model changes, approvals, context compression, task cancel, and renames. (#16225) by @doudouOUC

Documentation

• Release documentation now explains why publish can remain queued on GitHub-hosted runners and why those steps are not moved to ECS pools. (#16194) by @yiliang114
• Managed Agent structure-review documentation now reflects the completed store decomposition and current audit status. (#16230) by @wenshao

Internal Changes

• Adds an /autofix-pr chain test for rerunning a timed-out job when the rerun answer is lost. (#16138) by @wenshao
• Makes Local, SSH, Kubernetes, External, and Static runtime backends optional so embedded servers can include only needed backends; pending operations resume when available. (#16141) by @doudouOUC
• Simplifies managed-agent row locking and runtime verification while preserving session, turn, journal, and checkpoint validation order. (#16139) by @doudouOUC
• Refactors internal Managed Agent session-store consumers to narrower role interfaces without changing user-visible behavior. (#16121) by @wenshao
• Adds regression tests covering removal of active persistent permission rules, including post-removal decisions and preservation of unrelated rules. (#13818) by @TYY-2005
• Pins hook telemetry redaction tests with value-level assertions so future raw-text properties fail instead of leaking. (#11785) by @RunMintOn
• Reserves Hosted integration-test ports below ephemeral ranges to reduce port conflicts and improve test reliability. (#16174) by @wenshao
• Adds a regression test ensuring completed background shell notifications reach the next model request in the same turn without starting extra notification turns. (#16176) by @yiliang114
• ci(e2e): run the nightly suite with the unified agent loop on (#16192) by @qqqys
• Managed Agent's shared MySQL schema now follows documented conventions for identifiers, booleans, counters, indexes, and digests, improving SQL consistency. (#16143) by @doudouOUC
• EmbeddedRuntimeBroker now depends only on SessionQueryStore, simplifying Managed Agent internals without changing user-facing behavior. (#16216) by @wenshao
• Integration tests now install the local extension fixture from a dedicated subdirectory, fixing E2E failures caused by copying the extension store into itself. (#16222) by @yiliang114
• Improves managed-agent test reliability by refreshing Workspace owner membership expectations for concurrent sessions. (#16240) by @doudouOUC
• Fixes ACP parity tests by keeping session runtime state out of the compared workspace directory. (#16224) by @yiliang114

New Contributors

• @TYY-2005 made their first contribution in #13818
• @RunMintOn made their first contribution in #11785
• @Adkid-Zephyr made their first contribution in #15558

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.6...v0.26.7

0.26.6 - 2026-11-02

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• Adds a Web Shell sidebar activity view for trusted workspace conversations, showing approval/input/error states, unread completions, running work, and bulk read/archive controls. (#16022) by @deggs7
• Adds an "Add to Split View" option to eligible sidebar session menus in web shell, making it easier to open sessions side by side. (#15956) by @4ekuct25
• Adds request success rate, output speed, per-model success indicators, and distinct model colors to the Web Shell token usage panel. (#16108) by @qqqys
• Web Shell composer shows a dismissible Workflow chip that turns Workflow mode off, warns when tools are unavailable, and summarizes the mode in the + menu. (#16107) by @qqqys
• Web Shell Token Usage now plots per-request output speed, first-token latency, and failure details by model, making session performance and errors easier to inspect. (#16111) by @qqqys
• With experimental.unifiedAgentLoop enabled, more ACP turns, including channel prompts and Goal-proposal approvals, run on the shared loop. (#15902) by @qqqys
• Mods can now run prompt.submit hooks in headless runs for -p prompts and stream-json user messages, allowing prompt inspection or rewriting before submission. (#15905) by @qqqys
• Desktop releases now publish Fedora 44 and RHEL 10 x64 RPM installers alongside existing Linux packages. (#16060) by @yiliang114
• feat(web-shell): make GitHub channel configuration usable (#16096) by @wenshao
• feat(verify-pr): inherit --publish consent on same-session re-verification (#16112) by @wenshao
• feat(ci): let the PR review canary cohort carry its own endpoint (#16130) by @yiliang114

Bug Fixes

• OAuth refresh lock cleanup now removes the lock only when the current process owns it, preventing exiting processes from deleting active refresh locks. (#15661) by @2SinceJuly
• MCP tool discovery now skips duplicate tool names but keeps all uniquely named tools, preventing one name collision from hiding an entire server's tools. (#15757) by @2SinceJuly
• The Web Shell Plugins page now shows a back control when the navigation rail is unavailable, letting users return to chat from compact layouts. (#15651) by @2SinceJuly
• LSP code-navigation queries now fail with an explicit reason when no language server is ready instead of returning empty results. (#15946) by @404khai
• The verify report now counts targeted gate test-suite results separately from a round's scripted assertions, making assertion totals reflect only round checks. (#16013) by @wenshao
• Makes long /btw answers scroll in a bounded panel with mouse wheel, Shift+Up/Down, PageUp/PageDown, and a scrollbar, while short answers remain unchanged. (#15982) by @2SinceJuly
• Fixes verify-pr to count each suite run as one assertion and handle skip notices on closed PRs so review rounds are not rejected with zero assertions. (#16050) by @wenshao
• Fixes Kimi K3 generation recognition so supported IDs use the 1,000,000-token context window, while k3-256k keeps its 256K window. (#16048) by @wenshao
• Fixes review command so plain-language re-review requests target the intended PR and do not misread resume flags or file-like tokens. (#15998) by @wenshao
• Installs DejaVu fonts in the sandbox so capture cards use the intended monospace font and render legibly. (#16032) by @yiliang114
• Moves the Workflow mode toggle from the composer toolbar into the composer “+” menu to reduce toolbar clutter while keeping it accessible. (#16047) by @qqqys
• Improves managed-agent recovery accuracy by blocking gap fills without child ledger evidence and clarifying unknown or never-started states. (#16015) by @wenshao
• Fixes managed-agent child result delivery status so consumed child results are no longer marked pending. (#16074) by @wenshao
• Fixes AI attribution notes for loosely spelled git commit --amend commands so they diff against the amended commit, not earlier commits. (#16073) by @yiliang114
• Makes loop-protection reminder delivery and telemetry settle only when the model request carrying the reminder is accepted, preventing premature clearing or logging. (#15559) by @suzhiguo7
• fix(core): don't treat fd duplication (2>&1) as a file write in shell permission checks (#15681) by @Suaroman
• Sanitizes goal approval text and edit-diff file paths in Web Shell by escaping control and bidi characters so displayed approval content cannot carry hidden controls. (#16080) by @yiliang114
• Shows Managed Session list failures as a localized Retry notice with request details and avoids the empty state until loading succeeds. (#16082) by @yiliang114
• Fixes invalid Javadoc left after the module split so the managed-agent release profile can build documentation again. (#16099) by @doudouOUC
• Moves the Web Shell managed session list failure notice below the Sessions / Archived tabs and prevents its request id from being cut off. (#16104) by @yiliang114
• Fixes qwen-managed-agent-upgrade start and verify so they reach the control plane and Hosted Harness using the server's resolved endpoint configuration. (#16051) by @wenshao
• Tightens the Web Shell mobile composer layout by aligning workspace controls, truncating long names, and centering the context-usage ring. (#16103) by @ytahdn
• VS Code companion session history now groups This Week by the current calendar week, so earlier sessions are placed in Older instead of incorrectly appearing under This Week. (#16097) by @yiliang114
• The bundled review skill now tells fetch-pr to re-run with both --since and --since-model and removes the retired range model, preventing incomplete anchors. (#16094) by @yiliang114
• The Managed runtime release workflow can now rebuild the missing 0.26.5 runtime from its original tag after repairing the Javadoc issue that blocked the release. (#16114) by @doudouOUC
• Windows installer updates now stop the bundled runtime before replacing files, reducing update failures when node.exe remains locked. (#16071) by @yiliang114
• Channel users can now /approve or /deny their own pending one-time tool permission requests in the current attended turn. (#16087) by @BenGuanRan
• Linux AppImage builds now verify required dynamic libraries are bundled or documented as host-provided, helping prevent missing-library failures. (#16062) by @yiliang114
• Mods that register prompt.paint or ui.place hooks now load successfully, with those unsupported hooks skipped instead of stopping startup. (#16124) by @qqqys
• fix(mcp): warn about disabled tool typos on every discovery path (#16120) by @yiliang114
• fix(managed-agent): Restore profile-owned runtime verification (#16134) by @doudouOUC
• fix(core): attribute background shell notifications to the registering turn (#16125) by @yiliang114
• fix(core): report redirects on cd segments (#16129) by @yiliang114

Documentation

• Updates daemon usage documentation with current deployment, security, authenticated startup, token sidecar lifecycle, and lifecycle-command guidance. (#16056) by @wenshao
• Adds a README quickstart for daemon mode covering qwen serve startup, status/log/stop commands, default loopback binding, and local security warnings. (#16053) by @wenshao
• Updates managed-agent structure-review documentation to reflect the Runtime Broker extraction and newly separated recovery, reconciliation, and cancellation components. (#16109) by @wenshao
• ManagedAgentStore documentation now reflects the completed store conversion and removes outdated setter descriptions. (#16118) by @wenshao

Internal Changes

• Public and WebShell Session event streams now share replay and live delivery while keeping existing payloads, event names, and delivery order. (#15952) by @doudouOUC
• ci: route the classify_pr gate to ECS for every pull request (#16019) by @yiliang114
• Adds an end-to-end test for /autofix-pr watch helpers against real git and a fake gh to improve automated fix reliability. (#16017) by @wenshao
• Adds a managed-agent contract test asserting A3 sibling call binding and approval snapshot consistency, with no production behavior change. (#15995) by @wang1408
• Makes Windows preview installer tests run reliably through cmd.exe so nightly results are reported correctly. (#16028) by @yiliang114
• Improves managed-agent broker reliability by separating runtime execution workflows. (#16029) by @doudouOUC
• Makes managed-agent session store initialization safer by using constructor dependency injection. (#16061) by @doudouOUC
• test(managed-agent): gate the expired-cursor resync answer (#16076) by @wenshao
• Refactors Managed Agent runtime execution to centralize dispatch, cancellation, and completion tracking in an internal dispatcher. (#16084) by @doudouOUC
• Adds a Managed Agent test verifying mcp-release refusal and same-id retry behavior when a session uses two MCP servers. (#16089) by @wenshao
• Adds a managed-agent test verifying detach is refused while a tool result is unknown and the session remains attached. (#16098) by @wenshao
• Reconnects Hosted fault probes to current execution owners so managed-agent tests observe dispatch, cancellation, and delayed replies accurately. (#16101) by @doudouOUC
• Settings documentation now matches the runtime schema, and memory.enableAutoSkill and security.auth.useExternal are correctly shown as false. (#16085) by @yiliang114
• Windows installation PATH persistence is now protected by a contract test, reducing the chance that a fresh install fails to remain on PATH. (#16072) by @yiliang114
• Managed Agent storage was reorganized into Agent and Session stores without changing user-visible behavior. (#16090) by @doudouOUC
• refactor(managed-agent): trim store delegation boilerplate (#16142) by @doudouOUC

New Contributors

• @404khai made their first contribution in #15946
• @wang1408 made their first contribution in #15995
• @suzhiguo7 made their first contribution in #15559
• @Suaroman made their first contribution in #15681

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.5...v0.26.6

0.26.5 - 2026-11-01

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• Adds bounded SQL history queries for Managed local H2 and Hosted MySQL, enabling paged recent messages, journal transactions, and exact operation or command lookups. (#15839) by @doudouOUC
• Enables Generated UI by default and lets users continue chatting with interactive card values attached as removable context above the composer. (#15856) by @qqqys
• feat(web-shell): keyboard shortcuts for Session Overview and Split View (#14179) by @4ekuct25
• Adds a Russian locale for the Web Shell Goal card and Goals dialog, with English fallback for untranslated strings. (#13685) by @4ekuct25
• Extends Web Shell one-click daemon updates to pnpm, yarn, bun, and sudo-prefixed npm installs, and shows manual steps when automatic update is unavailable. (#15849) by @wenshao
• Web Shell and Qwen Code Desktop now show active background agent counts in the status-bar task pill while agents are running. (#14002) by @4ekuct25
• Managed WebShell now reads the latest 100 live items directly and pages upward for older messages, while live updates continue. (#15865) by @doudouOUC
• The /verify-pr local flow now reuses the current working tree only when it is a clean checkout of the PR head and records round details. (#15913) by @wenshao
• Archived conversations in Settings now show Export conversation and Restore inline beside More, making archive management faster. (#15872) by @deggs7
• PR watch now merges a stale base before attempting fixes when CI failures may be caused by the branch being behind. (#15922) by @wenshao
• Adds an experimental one-command Helm deployment for Managed Agent with Hosted Harness, MySQL, local execution, and an authenticated gateway. (#15492) by @doudouOUC
• Adds --in-place mode to the PR-watch skill to co-drive in your checkout, push the branch, and reply on the PR. (#15929) by @wenshao
• Keeps skill slash-command argument hints visible while typing and shrinks them to missing arguments. (#15930) by @wenshao
• Makes context memory files in the Context Usage panel clickable to open read-only previews such as AGENTS.md. (#15932) by @wenshao
• Makes autofix-pr post a pull-request comment when a round is blocked, stating what was attempted, what blocks it, and what human action is needed. (#15978) by @wenshao
• Adds a retry budget to /autofix-pr so a check left failing after three fix pushes is handed to a human and shown in /autofix-pr status. (#15985) by @wenshao
• Adds a Workflow switch in Web Shell, split panes, and VS Code chat to update saved session Workflow preferences for the next turn without interrupting responses. (#15779) by @qqqys
• Enables channel adapters to expose tools to agent sessions and adds QQ moderation tools, including recall, mute, kick, and join approval, controlled by moderationTools. (#14348) by @Eric-GoodBoy-Tech
• Adds a batch request and SDK method to fetch live-session snapshots for up to 20 selected workspaces in one call. (#12513) by @XIQIXIQIXIQI
• feat(cli): support exact-version updates without npm discovery (#11486) by @XIQIXIQIXIQI
• Improves the bundled code-review skill by re-anchoring after head drift, running actionlint checks, accounting for blockers, and refusing retries. (#16001) by @wenshao

Bug Fixes

• Hardens the managed-agent upgrade command's check, backup, install, start, and verify phases, including cleanup after interrupted installs. (#15853) by @wenshao
• TUI CI now runs impact-selected checks and reserves full source TUI acceptance for nightly runs, manual dispatch, and high-risk changes. (#15864) by @yiliang114
• Main pushes now run one macOS E2E shard instead of two, while schedules, manual dispatch, and feat/e2e branches keep the two-shard matrix. (#15873) by @yiliang114
• The Web Shell floating approval card for tool approvals now uses a single scroll region, avoiding nested scrollbars in long diffs. (#14621) by @4ekuct25
• Release builds and eligible downstream jobs now use labelled runners on existing ECS hosts to reduce hosted-runner queue delays. (#15832) by @yiliang114
• CUA input, window selection, and session cleanup align with 0.27.0 across CUA SDK, Browser SDK, and Node REPL for more reliable sessions. (#15784) by @LaZzyMan
• Review command tests now pin ghApiAllNested pagination and response merging behavior to prevent regressions in GitHub API requests. (#12560) by @indexGin
• The No-workspace sessions header now matches Pinned, Workspaces, and Recent with consistent height, muted text, spacing, chevron, and hover/focus behavior. (#15897) by @deggs7
• E2E builds now use the same pool-first runner routing and disk gate as other jobs, reducing hosted queue delays for post-merge test signals. (#15885) by @yiliang114
• Web Shell browser smoke tests now run on isolated ubuntu-latest runners to avoid shared-runner network failures. (#15886) by @yiliang114
• Web Shell now hides the No-workspace sessions sidebar section when it loads empty, matching Pinned behavior. (#15916) by @deggs7
• fix(web-shell): keep "+ Add session" clickable when the catalog is empty (#15835) by @2SinceJuly
• fix(ci): weight PR owner assignment by open issue load (#15194) by @DragonnZhang
• fix(cli): separate instruction files from auto memory in context usage (#15863) by @callmeYe
• fix(core): skip memory imports of files already loaded as context files (#15687) by @tneemo
• fix(core): retry bodyless 400 responses instead of treating them as fail-fast (#15821) by @Andrea-Bruno
• fix(cli): share hosted harness deadlines and bail after CI retries (#14752) by @zonemeen
• fix(web-shell): prune split panes whose session left the active catalog (#14577) by @4ekuct25
• fix(core): stop the session LSP client on Config shutdown (#14778) by @he-yufeng
• Session messaging guidance now explains safe-mode suppression or missing peer addresses without implying recipients do not exist. (#13948) by @haomint
• Memory writes now report omitted index entries and scan-cap omissions separately, with sample names and preserved warnings. (#14937) by @jabrailkhalil
• Web Shell tool-card diffs now wrap long lines so diffs with long comments or paths scroll vertically instead of showing per-line horizontal scrollbars. (#13272) by @ifan855
• /context now displays token counts of one million or more as millions in terminal, plain text, and OpenTUI views. (#13237) by @Shizoqua
• Fixed a CLI test-suite build failure by completing the settle-budget rename in hosted harness tests. (#15935) by @wenshao
• The sandboxed /verify-pr publisher now runs the artifact check itself and marks the round inconclusive if the check fails or cannot run. (#15938) by @wenshao
• Startup probe now separates broker write failures from wrong topic type, giving clearer fail-fast guidance and mqadmin remedy. (#15941) by @wenshao
• Fixes /verify-pr publishing to use the newest round and updates skill text to match CI harness-fault behavior. (#15966) by @wenshao
• Queues peer-controller registry writes from the same process before taking the cross-process file lock to reduce lock contention and make concurrent updates more reliable. (#15971) by @wenshao
• fix(managed-agent): stop a drained pool owner's not-found from retiring automations (#15990) by @wenshao
• Routes helper-test temporary files to the runner data-disk temp directory to prevent CI failures caused by a small host /tmp tmpfs. (#15979) by @yiliang114
• Exempts non-TUI package trees from the parity and no-flicker gates so PRs touching only them skip those lanes. (#15937) by @yiliang114
• Fixes session-commit registration when an absolute cd or pushd resolves to the same repository or a subdirectory, keeping commits tracked. (#15953) by @yiliang114
• isAoneHostFamily now uses the shared host normalizer so review remote matching and write routing judge Aone hosts consistently. (#15934) by @yiliang114
• fix(vscode): gate the companion vitest worker cap on the shared pool (#15940) by @yiliang114
• Peer-controller registry writers now wait up to 8s for a contended lock and can take over a crashed holder's lock without losing writes. (#15992) by @wenshao
• The CLI now requests consent for native extension commands before direct installation in OpenTUI; replacement or unmount cancels pending requests. (#15947) by @chiga0
• Fixes managed-agent cancellation, Maven embedding, Hosted fault testing, and SDK stream-idle recovery so verification flows behave reliably. (#15927) by @doudouOUC
• Makes each /verify round publish and update only its own status comment instead of overwriting other verification comments. (#15988) by @wenshao
• Fixes deep-verification reports so the verdict line stays a plain paragraph and the published comment is verified against the intended body. (#16002) by @wenshao
• Adds the missing development proxy for autofix-pr routes so the PR watches page reaches the daemon. (#16004) by @wenshao
• Updates MCP client and core to 2.4.0 to drop pre-bundled fast-uri 3.1.0 and improve dependency auditability. (#15925) by @yiliang114
• Uses the automatic GitHub job token for release publish checkout, reducing authentication failures while keeping PAT for branch push and release creation. (#15730) by @qwen-code-dev-bot
• Adds isolation so /verify-pr mutation matrix runs in its own tree and never on live measured cells. (#15999) by @wenshao
• Fixes /verify-pr PR badge binding when a flag is glued directly to the PR number, such as 15979--publish. (#15989) by @wenshao
• Extends built-in reasoning effort to DeepSeek-V4 and new Qwen prefixes while keeping explicit overrides. (#15987) by @wenshao
• Reports a registry that keeps changing during a write as PeerControllerError with code registry-changed and a retry message. (#16007) by @wenshao

Performance

• Release builds now reuse verified native audio prebuilds from matching successful main releases to skip redundant native builds. (#15854) by @yiliang114

Documentation

• Managed agent upgrade docs now state that SIGKILL skips cleanup and leaves a temporary upgrade directory that operators should remove. (#15878) by @wenshao
• Updated CI documentation to match the actual qwen-autofix job timeout and worst-case budget values. (#15914) by @yiliang114
• Adds a bilingual Managed Agent server configuration reference generated from server metadata to keep it current. (#15939) by @wenshao
• Adds a bilingual Managed Agent design document with a structure review and staged refactor proposal. (#15942) by @wenshao
• Reorganizes the bilingual Managed Agent deployment guide into three tiers for first-time deployments. (#15944) by @wenshao
• Adds documentation proposing scheduled /autofix-pr --in-place rounds that report findings and require explicit approval before editing. (#15960) by @wenshao
• Updates Managed Agent server docs to link setting tables to the generated configuration reference instead of duplicating settings. (#15965) by @wenshao
• Clarifies artifacts.read-timeout documentation to qualify storage stall behavior and retain the per-row lease note. (#15980) by @wenshao
• Managed-agent docs now note that private CSI sessions record no committed tool receipts and correct lease wording for artifacts.enabled. (#16000) by @wenshao
• Updates managed-agent documentation to describe extracted collaborators and reconciliation prerequisites without changing behavior. (#16012) by @wenshao
• Documents the ManagedAgentStore setter investigation and its Spring injection findings. (#16014) by @wenshao

Internal Changes

• Updates Web Shell smoke tests to use the renamed Add workspace dialog and Workspaces sidebar label. (#15860) by @wenshao
• Adds a regression gate for managed-agent failover that verifies recovery when a replacement host loses its first takeover-load reply. (#15857) by @wenshao
• Managed public items, transcripts, recent history, and export now use bounded current-message paging instead of persisted snapshots. (#15870) by @doudouOUC
• Hosted fault tests now use Managed JDBC stores so identity handling matches production and avoids false test failures. (#15882) by @yiliang114
• Web shell test setup now selects the retained radio via keyboard focus and Space, making Generated UI retention checks more stable. (#15896) by @yiliang114
• Hosted recovery and transcript fixtures now follow current cancellation, cleanup, and transcript-event contracts for more reliable managed recovery tests. (#15899) by @yiliang114
• Web shell recovery tests now include transcript coverage and temporary storage routing so recovered sessions show confirmed answers and latest requests. (#15883) by @yiliang114
• Improved managed-agent test reliability by retrying expected 503 projection-pending reads during coordinator restart checks. (#15910) by @yiliang114
• Managed tests now count cancellation takeovers only after successful HTTP acceptance while recording all load attempts. (#15912) by @yiliang114
• Managed Agent Server is now organized into Maven modules while remaining a single executable service. (#15871) by @doudouOUC
• Managed crash-recovery validation now treats targeted cleanup loads as non-resuming recovery while still requiring cancellation takeover. (#15926) by @yiliang114
• Adds a failover test verifying Managed Agent recovery when a generation crashes before committing assistant output. (#15881) by @wenshao
• Managed Agent session creation now uses a single SessionCreationCommand and fails closed when workspace creation is unsupported. (#15955) by @wenshao
• Adds an E2E fault test ensuring Managed Agent recovers correctly when Harness crashes between sibling approvals. (#15945) by @wenshao
• Restores classifier-driven web-shell smoke routing to self-hosted ARM64 for admitted runs, with ubuntu-latest fallback for fork PRs. (#15961) by @yiliang114
• Adds test evidence and documentation pins for the A3 original-plan count and corrects related prose without changing production behavior. (#15977) by @wenshao
• Splits AgentStateStore into six role interfaces so managed-agent components can depend on narrower session-state contracts without behavior changes. (#15962) by @wenshao
• Tightens CI checks so the Runtime Broker fault gate only relies on Hosted process setup steps that cannot be skipped or ignored after failure. (#15964) by @wenshao
• Adds a CI fault scenario that crashes the managed-agent harness after earlier tool rounds to verify recovery of round, cost, and token state. (#15981) by @wenshao
• Adds tests confirming Aliyun OSS tool-publication object open operations retry transient service errors and client timeouts before succeeding. (#15983) by @wenshao
• Refactors Runtime Broker by extracting the UNKNOWN-execution lookup cooldown into a dedicated component without changing behavior. (#15986) by @wenshao
• Routes the Lint & Static lane to self-hosted ECS runners for all pull requests, including forks, while preserving timeout and override controls. (#15974) by @yiliang114
• Fixes Hosted attachment and harness-client lookups for decorated HarnessConnector beans by resolving the underlying connector state. (#15984) by @wenshao
• qwen review drive now builds its not-started unavailable report from one factory, keeping zero values and notes unchanged. (#15943) by @yiliang114
• Managed-agent recovery tests now verify B2 replacement attempt and stream identities, blanked retracted prefixes, and the live SSE view. (#15993) by @wenshao
• The review-salvage replay test now derives the supersede kill-dir prefix from the workflow text, preventing the stub glob from drifting. (#15963) by @yiliang114
• Refactors managed-agent runtime hold inventory handling without changing user-visible behavior. (#15997) by @wenshao
• Adds tests pinning private CSI receipt guards and applies a follow-up managed-agent design correction. (#16011) by @wenshao
• Adds tests covering managed-agent lapsed-lease ledger writes and result-store wiring, with no direct user-facing change. (#16018) by @wenshao

New Contributors

• @indexGin made their first contribution in #12560
• @tneemo made their first contribution in #15687
• @ifan855 made their first contribution in #13272

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.4...v0.26.5

0.26.4 - 2026-10-29

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• feat(generated-ui): follow the host theme in generated views (#15798) by @qqqys
• feat(ci): bound autofix takeover with a round budget (#15388) by @wenshao
• feat(verify-pr): make "re-verify after update" a follow-up trigger (#15816) by @wenshao
• feat(extensions): support unscoped npm installation and updates (#15781) by @callmeYe
• feat(web-shell): move no-workspace sessions beside workspaces (#15319) by @deggs7
• feat(managed): add independent worktrees for main sessions (#15745) by @doudouOUC
• feat(web-shell): allow embedding hosts to pick workspace directories (#15296) by @jkxiao
• feat(generated-ui): localize and reshape the card frame (#15803) by @qqqys
• feat(web-shell): show model timing inline in Token Usage (#15774) by @qqqys
• feat(cli): raise workspace skill install limits (#15030) by @callmeYe
• feat(dws): configure quoted replies by conversation type (#15019) by @qqqys
• feat(web-shell): present Generated UI cards as interactive views (#15831) by @qqqys
• feat(autofix-pr): root watches at the session worktree under the daemon (#15817) by @wenshao
• feat(managed-agent): add a phased upgrade command to the reference distribution (#15792) by @wenshao
• feat(desktop): extend the PATH of the daemon the Desktop spawns on macOS (#15190) by @4ekuct25
• feat(verify-pr): stamp model, CLI and skill version on published reports (#15838) by @wenshao
• feat(cua): enable default cursor and live macOS PiP (#15669) by @LaZzyMan
• feat(generated-ui): offer interactive views when users want to adjust values (#15846) by @qqqys
• feat(acp): run the stop point of fresh prompts on the client's chain (#15695) by @qqqys

Bug Fixes

• fix(config): report controls that were silently dropped (#15794) by @yiliang114
• fix(cli): probe ICU before the UI import so an ICU-less Node fails loud (#11753) by @he-yufeng
• fix(core): keep reasoning and function_call items adjacent through Responses cleanup (#11684) by @he-yufeng
• fix(cli): carry branch points through the session load replay (#13814) by @he-yufeng
• fix(web-shell): pause requests offline and recover automatically (#14980) by @ytahdn
• fix(cli): map held pidfile lock and identity refusals to one stderr line (#15818) by @yiliang114
• fix(release): unblock Managed runtime 0.26.3 backfill (#15825) by @yiliang114
• fix(cli): prevent provider menu overlap in short terminals (#15807) by @yiliang114
• fix(core): infer DeepSeek Flash effort on Alibaba proxies (#15828) by @callmeYe
• fix(ci): exempt surefire-bound IT classes from the failsafe roster (#15830) by @yiliang114
• fix(managed-agent): retry owed Hosted cleanup after its generation dies (#15824) by @wenshao
• fix(verify-pr): bind reports to their round snapshot (#15746) by @jianyx1
• fix(web-shell): standardize workspace and session display copy (#15315) by @deggs7
• fix(web-shell): list and open standalone sessions in the Session Overview (#13679) by @4ekuct25
• fix(cli): limit provider updates to the models served to the stored key (#15230) by @qqqys
• fix(cli): return clean status for confirmed OpenTUI keyboard quits (#15723) by @chiga0
• fix(web-shell): preserve model names after selection (#15841) by @callmeYe
• fix(acp): keep setCoreValue resolving when the output-language write fails (#15848) by @yiliang114

Performance

• perf(web-templates): reclaim 92 KB of the HTML export renderer bundle (#15778) by @yiliang114
• perf(managed): Bound activation renewal journal growth (#15813) by @doudouOUC

Documentation

• docs(managed): cite the shared-Session queue decision in the gate design (#15797) by @wenshao

Internal Changes

• test(managed-agent): pin HostedProcessLiveness.running on every host (#15791) by @wenshao
• test: settle the two loaded-host waits from #15593 (#15639) by @2SinceJuly
• test(core,cli): stop flat 30s budgets shadowing the raised ECS timeout ceiling (#15801) by @yiliang114
• test(managed-agent): qualify test fixtures onto the managed JdbcTemplate (#15809) by @yiliang114
• ci(java): remove MariaDB full integration lane (#15827) by @doudouOUC
• test(managed-agent): bind the full-application suites to the managed beans (#15802) by @yiliang114
• test(managed-agent): verify journal constraint rollback across databases (#15842) by @yiliang114

New Contributors

• @2SinceJuly made their first contribution in #15639
• @jianyx1 made their first contribution in #15746

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.3...v0.26.4

0.26.3 - 2026-10-28

Highlights

• Generated UI cards keep inputs and state across navigation and split views, show live views after initial calculation, and appear only for clients that can render MCP Apps. (#15611, #15787, #15720, #15790)
• Persistent session Workflow mode can be enabled with --workflow-mode on, /workflows mode on, SDK, or ACP, and workflow switches now reach the CLI sandbox. (#15422, #15743)
• Tool results shrink to remaining headroom before auto-compaction, and qwen-latest-series- or qwen4 aliases are detected with 1M-token context windows. (#13599, #15758, #15762)
• Valid XML tool calls are recovered when parameters quote other call markup, malformed blocks are isolated, and quoted examples are not dispatched. (#13579)
• ACP and daemon clients keep rewind identity consistent across retries, cancellations, checkpoints, recordings, and notifications, while SDK frames preserve subagent identity. (#15441, #15657)
• /autofix-pr start stops when the PR head branch has unpushed commits or uncommitted changes, preventing the watch from overwriting unseen work. (#15736)

Breaking Changes

No known breaking changes.

Complete Change List

Features

• feat(memory): opt-in extraction cadence after a no-op run (#13571) by @yiliang114
• Shrinks tool results to remaining headroom before auto-compaction, respecting explicit truncation settings and protected content to reduce unnecessary compaction. (#13599) by @yiliang114
• Keeps eligible Generated UI cards alive across history navigation, session switching, and split-view changes in the same page, preserving inputs, controls, and runtime state. (#15611) by @qqqys
• Adds persistent session Workflow mode through --workflow-mode on, /workflows mode on, SDK, and ACP controls, with saved mode shown in Ink and OpenTUI. (#15422) by @qqqys
• Adds opt-in composerVariant simple to Web Shell, providing a compact input with a limited add menu and shared editor, attachments, and session actions. (#15663) by @samuelhsin
• /autofix-pr start now stops if the PR head branch has unpushed commits or uncommitted changes, preventing the watch from overwriting unseen work. (#15736) by @wenshao
• Adds administrator-driven reclamation and reuse of idle resident CSI Workspaces with cleanup verification and atomic release. (#15729) by @doudouOUC
• OpenTUI terminals now draw the Mods band above the prompt and run Mod ui.render hooks there, matching the Ink terminal experience. (#15699) by @qqqys
• ACP live frames, persisted replays, SDK normalization, and transcript reduction now preserve subagent identity for thoughts, messages, and tool frames. (#15657) by @yiliang114
• Archived conversations are now managed in Settings → Archived conversations in Web Shell, keeping archive groups, filters, and lifecycle actions. (#15557) by @deggs7
• Managed Agent can now be embedded opt-in in a Java host's JVM and Spring context, with dedicated connection pooling or a supplied DataSource. (#15718) by @yiliang114
• /verify-pr now supports a failures ledger and rejects findings or blocked verdicts unless failures are attributed to the PR. (#15764) by @wenshao
• Adds the qwen-managed-agent-admin launcher for running managed-agent offline maintenance commands by name, with help and unknown-command handling. (#15772) by @wenshao
• feat(generated-ui): check a view's initial calculation before display (#15790) by @qqqys

Bug Fixes

• Recovers valid XML tool calls when parameters quote other call markup, isolates malformed blocks, and prevents quoted examples from being dispatched. (#13579) by @yiliang114
• Carries rewind identity through ACP/daemon clients so retries, cancellations, checkpoint resolution, recording, and notifications stay consistent. (#15441) by @yiliang114
• Fixes release publishing checkout failures by using the run-scoped GitHub token for read-only checkout while keeping the bot PAT for write steps. (#15728) by @qwen-code-dev-bot
• Pins pull request CI retries to the head SHA from the triggering event so reruns test the intended commit. (#15734) by @yiliang114
• fix(cli): forward workflow switches into the CLI sandbox (#15740) (#15743) by @qwen-code-dev-bot
• Recognizes invite-only qwen-latest-series- model aliases as 1M-context models in the context-window detector. (#15758) by @wenshao
• Makes daemon-managed channels honor approvalMode even when no webhooks are configured, matching standalone startup. (#15754) by @qqqys
• Shows the exception class in managed-agent scan summaries for non-first row failures that have no message. (#15725) by @yiliang114
• Workspace Skill inventory queries now return a retained snapshot immediately and refresh in the background, while cold reads, deletion, and enablement wait for fresh data. (#14890) by @callmeYe
• qwen4 model identifiers are now sized with a 1,000,000-token context window by default when no upstream catalog entry defines them. (#15762) by @wenshao
• Acceptance tests now wait for the focused shell Tab byte to reach the child process before asserting exact input bytes, preventing flaky failures. (#15763) by @yiliang114
• Managed runtime distribution now publishes verified server artifacts to GitHub Releases and OSS without requiring Maven Central, keeping CLI downloads unchanged. (#15782) by @doudouOUC
• Generated UI is now offered only to sessions whose client can render MCP Apps, while unsupported clients receive a clear unsupported-client reason. (#15720) by @qqqys
• In Web Shell, live Generated UI cards now attach and show their interactive view immediately when tool results arrive instead of requiring reopen. (#15787) by @qqqys
• Desktop release preparation now resets Git metadata on self-hosted runners before checkout to avoid stale-workspace failures. (#15761) by @yiliang114

Documentation

• Documents managed-agent shared-Session queue behavior and advises giving slow-probed gated definitions their own Session. (#15724) by @yiliang114

Internal Changes

• Internal tests now compare git worktree prune dry-run results as an unordered set, making them insensitive to git output order. (#15767) by @yiliang114
• Managed snapshot pool test setup now allows a three-second database connection lease so schema creation on busy CI runners is less likely to fail. (#15749) by @yiliang114
• Installed npm and private Bun acceptance tests now run nightly and by manual dispatch instead of on every PR update, while source TUI PR behavior stays unchanged. (#15719) by @yiliang114
• Improved Managed Agent reliability testing by adding terminal recovery fault gates for failover scenarios. (#15756) by @wenshao
• Improved CLI test stability for review cleanup orphan-process handling by making process liveness deterministic. (#15751) by @yiliang114
• Hosted gates now use a shared zombie-aware liveness check, improving consistency when waiting for processes to stop. (#15777) by @wenshao

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.2...v0.26.3

0.26.2 - 2026-10-27

Highlights

See the complete change list below.

Breaking Changes

• fix(managed-agent)!: simplify Web Shell controls and retire cwd submissions (#15616) by @doudouOUC

Complete Change List

Features

• feat(managed): add session permission rules and Auto approval (#15448) by @doudouOUC
• feat(core): add an experimental shared agent loop and run headless on it (#15369) by @qqqys
• Adds X-DashScope-Source-Config to matching QwenCloud and Qianwen AI Platform requests to identify site and client, without changing authentication or retries. (#15041) by @yiliang114
• Resumed OpenTUI sessions now retain recorded text, images, tool results, and prompt identities, while rewind restores transcript, request, and draft. (#15384) by @chiga0
• OpenTUI now captures complete checkpoints at edit approvals and restores them safely, preserving conversation metadata and blocking stale or cancelled owners. (#15389) by @chiga0
• Adds automation.scanner-threads to run managed automation scanner rows on a dedicated pool and leases rows from claim time. (#15442) by @yiliang114
• Adds Continue analysis with current parameters so Generated UI users can review and send frozen parameters to the original session. (#15333) by @qqqys
• Adds an opt-in external-scratch Runtime provider that lets an operator-run HTTP service place Session workers while preserving attestation and recovery. (#15261) by @yiliang114
• Adds a prompt argument hint for /verify-pr so users can see the PR number and supported flags such as --publish and --watch. (#15463) by @wenshao
• Adds a durable queue for messages sent during a Managed Turn, allowing queued text and file references to be delivered, edited, withdrawn, or resumed safely. (#15449) by @doudouOUC
• Adds /autofix-pr status to list all watches in a clone and lets users resume a watch in a later session. (#15510) by @wenshao
• Adds a Git URL source to the Web Shell Add-workspace dialog so users can clone a repository and open it as a workspace in the browser. (#15361) by @wenshao
• Enables image and PDF delivery for managed runtime providers with a 768 KiB base64 media limit and a 896 KiB tool-result limit. (#13606) by @water-in-stone
• feat(managed-agent): give the Hosted Harness its Broker through the environment (#15244) by @wenshao
• Adds an operator endpoint to register immutable, idempotent EXTERNAL Workspace storage ownership with provider identity and tenant alias. (#15516) by @yiliang114
• UserPromptSubmit hooks no longer re-append byte-identical additionalContext within a session, avoiding duplicate context and keeping transcripts accurate. (#14129) by @yiliang114
• Mods can now observe turn.start and turn.complete events in headless and terminal runs, and use $.turn.abort to end a running turn. (#15443) by @qqqys
• Web Shell Settings now lets you list, add, and remove extra workspace directories, applying changes to live sessions and informing the model of new roots. (#15393) by @wenshao
• feat(autofix-pr): run the helpers through a tool, and let a confirmed watch's fixes run unattended (#15548) by @wenshao
• feat(managed-agent): let the bootstrap set the service account and MySQL up (#15543) by @wenshao
• feat(managed-agent): add pinned extension package management (#15490) by @doudouOUC
• feat(managed-agent): support Workspace file uploads (#15586) by @doudouOUC
• feat(core): Add durable terminal identity lookup (#15566) by @doudouOUC
• When experimental.unifiedAgentLoop is enabled, fresh ACP user prompts run their model round on the shared agent loop while ACP keeps its own tool-step handling. (#15509) by @qqqys
• The Mod runtime adds ui.render, .ui.resolve, and .ui.invalidate support, enabling render-site hooks and terminal element resolution before hosts draw UI. (#15591) by @qqqys
• feat(web-shell): add plain-text and Markdown reply copying (#15345) by @ytahdn
• feat(cli): connect OpenTUI voice input to the shared pipeline (#15617) by @chiga0
• feat(ci): canary-split the PR review model by PR number (#15538) by @yiliang114
• feat(daemon): add cached workspace artifact queries and batch removal (#15594) by @BZ-D
• feat(managed-agent): add native snapshots and HTTP recovery (#15569) by @doudouOUC
• feat(cli): admit Managed Runtime boot-v2 with external storage and publication surface in container mode (#15522) by @yiliang114
• feat(tools): let an explicitly configured search backend name its host (#15638) by @yiliang114
• feat(cli): align OpenTUI foreground shell controls with Ink (#15642) by @chiga0
• feat(acp): run the tool batches of fresh prompts on the shared agent loop (#15627) by @qqqys
• feat(serve): allow secondary sessions to cd within their workspace (#15323) by @rockybot2026
• feat(mods): draw the band above the prompt in the Ink terminal (#15619) by @qqqys
• feat(autofix-pr): run an unattended watch's fix agent in the background (#15578) by @wenshao
• feat(core): add local managed journal snapshot recovery (#15626) by @doudouOUC
• feat(verify-pr): add a local preflight and a verdict check (#15614) by @wenshao
• feat(web-shell): show /autofix-pr watches and stop them from the Web Shell (#15668) by @wenshao
• feat(mods): compile JSX in .tsx and .jsx Mod modules (#15624) by @qqqys
• feat(cli): bind the PR named by a /verify-pr prompt (#15625) by @wenshao
• feat(cli): complete OpenTUI extension discovery and source management (#15671) by @chiga0
• feat(managed-agent): add the external-workspace provisioning prerequisite (#15512) (#15654) by @yiliang114
• feat(web-shell): nest derived sessions under their parent in the sidebar (#15714) by @wenshao
• feat(web-shell): name new sessions from the first prompt (#15689) by @wenshao
• feat(managed-agent): check Harness compatibility, leftover workers and a live service in preflight (#15564) by @wenshao
• feat(web-shell): stream BTW answers with reconnect and scroll following (#15709) by @ytahdn
• feat(cli): run only the tests related to the diff in review build-test (#15547) by @wenshao

Bug Fixes

• fix(cli): harden serve token resolution against project env injection (#15228) by @yiliang114
• fix(cli): align OpenTUI cancellation and retry with Ink (#15283) by @chiga0
• test(managed-agent): assert web_fetch in the Hosted workspace gate fixture (#15451) by @yiliang114
• fix(vscode): leave edit mode after a refused edit (#15439) by @yiliang114
• fix(rewind): resolve legacy turns by source identity (#13737) by @yiliang114
• Retires registered Linux zombie workers before startup waits, preventing a 30-second delay and returning a prompt terminal conflict instead. (#15459) by @doudouOUC
• Preserves arrival order of direct-mode stream-json startup messages so pipelined client requests execute and return results in the intended order. (#15378) by @yiliang114
• Restores managed-agent module and CLI build lanes by correcting the schema inventory column total after new permission-rule columns were added. (#15455) by @yiliang114
• Makes the team-lead Hosted MySQL integration boot retry when the harness loses its reserved port, reducing flaky test timeouts. (#15224) by @qwen-code-dev-bot
• CLI checkpoint restores now preserve ownership, Read permissions, and prompt history so cold resume and later rewind behave consistently. (#15386) by @chiga0
• Eligible npm sessions using OpenTUI now run on private Bun 1.4.4 while preserving Node for npm subprocesses and existing encrypted settings. (#15387) by @chiga0
• The main CI failure reporter now files issues even when a newer main run is still pending, adding a provisional note instead of dropping the failure. (#15471) by @yiliang114
• Web Shell permission-rule declarations are now self-contained so package publishing no longer fails on unresolvable core-module imports. (#15501) by @yiliang114
• CLI preserves OpenTUI clipboard source attribution and collapsed large pastes during draft recovery, queue restoration, cancellation, and recall. (#15396) by @chiga0
• Hosted MySQL test isolation preserves the original table-truncation error when foreign-key restore also fails. (#15505) by @yiliang114
• CLI cost estimates no longer double-count reasoning tokens for OpenAI Chat Completions, OpenAI Responses, and Qwen OAuth usage. (#15464) by @yiliang114
• Release automation limits release-note summary requests to three concurrent waves, keeps result order, and refreshes release PRs from main before automated approvals. (#15500) by @yiliang114
• Managed Web Shell prevents duplicate message composers and keeps drafts stable during Session polling or permission-rule rerenders. (#15496) by @doudouOUC
• On Linux, Web Shell no longer shows the native directory Browse button when the X11 display is unreachable. (#15434) by @yiliang114
• Raises timeout budgets for the standalone Node/Bun regression suite on the release pool and cleans up outdated exemptions. (#15102) by @qwen-code-dev-bot
• Refuses to start pull-request autofix watch from a checkout that cannot run it and reports the required working directory. (#15506) by @wenshao
• Simplifies Web Shell status display when loading phrases are hidden and improves status text contrast. (#15423) by @deggs7
• Keeps incremental lint history fetching lightweight and surfaces fallback as a CI warning and step summary. (#15511) by @yiliang114
• Gives eligible main push SHAs separate CI concurrency groups so independent main validations are not displaced by newer pending runs. (#15508) by @yiliang114
• Authenticates managed-runtime OSS preflight checks so release publishing can verify existing artifacts and avoid failed uploads. (#15504) by @yiliang114
• Runs the self-hosted SDK Java test suite before taking the host lock to reduce timeouts and lock contention in CI. (#15254) by @qwen-code-dev-bot
• Prevents the main CI failure tracker from filing issues for post-merge PR reruns that do not reflect the merged tree. (#15250) by @qwen-code-dev-bot
• Hides unavailable Managed background tools so Background Shell and Monitor appear only when the session can run them. (#15497) by @doudouOUC
• Fixes a managed-agent shutdown error caused by lifecycle renewal scheduling. (#15231) by @qwen-code-dev-bot
• Unblocks independent same-tenant resident workspaces blocked by an unreclaimed binding and fixes hosted file tool admission. (#15457) by @doudouOUC
• In Web Shell standalone sessions, Language: UI is now a user setting, so changing it no longer fails with workspace-scope errors. (#15521) by @wenshao
• Failed queued automatic memory extraction now logs a debug warning instead of causing an unhandled rejection. (#15338) by @wenshao
• Credential-free release preparation jobs now run on shared ECS runners in the official repository, with existing fallbacks preserved. (#15517) by @yiliang114
• Headless CLI retries now show a retry classification on stderr, making transport failures and restart reasons easier to diagnose. (#15515) by @yiliang114
• In Web Shell, /autofix-pr watches now persist as durable tasks, so they survive idle session reclamation and appear in Scheduled tasks. (#15523) by @wenshao
• Deletes a turn's memory catalog snapshot record when the turn is removed, preventing orphaned snapshots while preserving catalog continuity. (#15540) by @yiliang114
• Fixes ARM64 CI lane pricing so self-hosted lanes get correct ceilings: lint 45 to 90 minutes and no-AK integration 30 to 60 minutes. (#15529) by @yiliang114
• Hosted sessions now persist Todo and project committed task lists, showing them in Managed WebShell so later turns use the latest state. (#15155) by @yiliang114
• Incremental lint CI now deepens history until the merge base is provable, reducing unnecessary full-repository scans on long-lived pull requests. (#15525) by @yiliang114
• Incremental Prettier checks now group files by the nearest configuration directory, so formatting rules and ignores apply correctly across touched paths. (#15531) by @yiliang114
• Kubernetes Harness pool now passes qwen serve the managed runtime broker URL and token environment pair, with clearer errors for missing, invalid, or non-HTTPS Broker URLs. (#15394) by @wenshao
• The tui-parity Check TUI impact gate now runs on the persistent ECS pool for trusted PRs, while untrusted fork PRs stay on GitHub-hosted runners. (#15561) by @yiliang114
• UserPromptSubmit hook context is no longer duplicated on the next turn after automatic compaction, and compression checkpoints are protected from later rescue mutations. (#15575) by @yiliang114
• The Mod runtime now targets Claude Code 2.1.310, accepts tool.check answers with setting fields, and reports that version from $.session.version(). (#15560) by @qqqys
• UserPromptSubmit hook context remains deduplicated after memory catalog snapshots or added message parts, preventing repeated hook context copies in CLI requests. (#15556) by @yiliang114
• fix(ci): resolve the pool's Maven dependencies through a domestic mirror (#15610) by @yiliang114
• fix(ci): stop failing the Maven bootstrap on a rate-limited pool egress (#15539) by @yiliang114
• fix(managed-agent): retry read-only attestation after EOF (#15608) by @doudouOUC
• fix(ci): resolve libfaketime for the host architecture (#15609) by @yiliang114
• fix(ci): bound setup-java on the pool and fall back to the tool cache (#15605) by @yiliang114
• fix(web-shell): match cancellation duration to turn summary (#15615) by @ytahdn
• fix(web-shell): hide turn navigation without sufficient gutters (#15604) by @ytahdn
• fix(core): stop warning about a placeholder in the attribution header (#15636) by @yiliang114
• fix(ci): resolve the pool's Maven dependencies from the intranet mirror (#15641) by @yiliang114
• fix: preserve interleaved reasoning in TUI and Chat replay (#15621) by @DragonnZhang
• fix(cli): preserve previews and verify installed OpenTUI distribution (#15655) by @chiga0
• fix(ci): Isolate Managed runtime release test classes (#15676) by @doudouOUC
• fix(ci): deepen until the lint scope's own acceptance holds (#15644) by @yiliang114
• fix(release): extend standalone runtime download retries (#15673) by @qwen-code-dev-bot
• fix(ci): preserve setup-java JDKs during pool fallback (#15623) by @yiliang114
• fix(ci): keep the sdk-java cache write-back on a failed suite (#15555) by @wenshao
• fix(serve): Fix Hosted local state and foreground Shell validation (#15670) by @doudouOUC
• fix(release): close the failure issue once a re-run passes (#15711) by @yiliang114
• fix(ci): refresh the stale ci.yml workflow-size baseline (#15717) by @wenshao
• fix(web-shell): keep session rename responsive (#15688) by @wenshao
• fix(cli): preserve OpenTUI extension update state (#15716) by @chiga0
• fix(ci): Preserve canonical artifacts across job retries (#15713) by @doudouOUC

Performance

• Runs web-shell and standalone tests concurrently with the main CLI test suite to shorten required Linux CI time. (#15438) by @wenshao
• Keeps unchanged legacy memory catalog snapshots stable in history to improve DashScope cache reuse, resume, and retry behavior. (#15246) by @yiliang114

Documentation

• Updates scheduled-task gate examples so copied first-run commands include required repository and PR options. (#15452) by @sxl-lucky7
• docs(design): external Hosted Workspace profile (phase 2) (#15520) by @yiliang114
• docs(managed): describe gate scanner capacity under the row pool (#15571) by @wenshao
• docs: clarify usage statistics across authentication methods (#15554) by @sxl-lucky7
• docs(qwencloud): document the forced attribution header and pin the admission split (#15633) by @yiliang114

Internal Changes

• test(cli): cover remaining Managed Workspace activation refusals (#15460) by @doudouOUC
• Corrects the managed-agent CSI inventory column count so test suites pass after permission-rule columns were added. (#15469) by @yiliang114
• test(ci): pin the disk-pressure collector revision and FULL_PAYLOAD run bodies (#15475) by @yiliang114
• Adds an integration test verifying the removeScratchDir retry budget removes busy scratch directories without logging cleanup warnings. (#15488) by @yiliang114
• Tests now capture expected disk-full log output so real CI disk-full failures are easier to distinguish from deliberate test errors. (#8549) by @EthDing
• Managed Agent tests now wait for a dead worker to become unusable before checking lease identity reuse or reconciling its loss. (#15503) by @yiliang114
• SDK Java CI retains Surefire and Failsafe failure reports and preserves original SQL causes in channel ownership race tests. (#14910) by @Aling33
• Adds CI regression coverage to keep JUnit report paths where CI collects them. (#15502) by @barry166
• Applies the slow-host policy to hosted approval test waits so slow runners get longer approval-test timeouts. (#15183) by @qwen-code-dev-bot
• Adjusts hosted approval test waits based on runner load policies to improve CI reliability on slower hosts. (#15101) by @qwen-code-dev-bot
• Adds automated checks for memory recovery and hard-rescue catalog accounting to prevent regressions. (#15535) by @yiliang114
• Adds automated checks for managed-agent row pool fan-out and corrects design documentation on shutdown and task limits. (#15528) by @yiliang114
• Adds a repeatable Harness Kubernetes qualification workflow for deploying, verifying, and exercising shared pools. (#15519) by @doudouOUC
• test(cli): pin the lifecycle-protocol refusal on the load route (#15536) by @yiliang114
• Adds tests verifying the lint --setup guard installs only required linters and installs nothing when no lane is selected. (#15533) by @yiliang114
• Consolidates catalog visibility rules so curated history reads consistently hide catalog snapshots when legacy memory is suppressed. (#15534) by @yiliang114
• Centralizes QwenCloud MaaS host list usage for telemetry classification and DashScope admission to keep host handling consistent. (#15532) by @yiliang114
• Hosted harness tests now use longer adaptive wait budgets for interrupted-model checks, reducing flaky timeouts without changing asserted behavior. (#15551) by @yiliang114
• test(managed-agent): align S1 acceptance fixtures with runtime contracts (#15585) by @doudouOUC
• Managed Agent workspace upload admission tests now use valid binary upload parameters and upload-status queries, and expect precise not-found errors for unknown uploads. (#15588) by @doudouOUC
• test(runtime-broker): build the extension-profile fixture paths portably (#15598) by @yiliang114
• test(cli): budget the skill-recovery poll through the settle policy (#15589) by @yiliang114
• test(lint): pin the .gitignore grouping arm and the batch stride (#15640) by @yiliang114
• Pin unknown sed long options (#15223) by @waybarrios
• ci: run every suite but CLI and core in the side test lane (#15570) by @wenshao
• test(cli): read xn-- hosts by the running URL parser in the Broker URL fixture (#15544) by @wenshao
• test(integration): widen the ready-prompt budget on GitHub-hosted macOS x64 (#15692) by @yiliang114
• chore(goal): remove an unused evidence helper and stale /goal hook comments (#15693) by @qqqys
• test(web-shell): raise the 200-record history-viewport smoke budget to 300s (#15656) by @yiliang114
• refactor(managed-agent): close the scanner row-pool residues (#15527) (#15632) by @yiliang114
• test(managed-agent): initialize the snapshot pool connection synchronously (#15715) by @doudouOUC
• test(cli): close the deferred review Suggestions from #15625 (#15710) by @wenshao

New Contributors

• @sxl-lucky7 made their first contribution in #15452
• @EthDing made their first contribution in #8549
• @Aling33 made their first contribution in #14910

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.1...v0.26.2

0.26.1 - 2026-10-24

Highlights

See the complete change list below.

Breaking Changes

• fix(qqbot)!: restore per-group session isolation under thread scope (#13882) by @Eric-GoodBoy-Tech

Complete Change List

Features

• Adds opt-in daemon token persistence: qwen daemon start --save-token stores the bearer token in an owner-only file, qwen daemon token manages it, and daemon stop removes it. (#15023) by @wenshao
• Adds idle-only manual context compression for Workspace-bound Managed Sessions via public API and the WebShell context-usage popover, preserving audit history. (#15104) by @doudouOUC
• Aligns OpenTUI input, editing, drafts, and /statusline behavior with Ink, adding configurable status lines and more predictable completion, queue, and cancel handling. (#14964) by @chiga0
• feat(shell): remind the model to record files a command writes (#15055) by @zjgzx1988
• Managed Workspace recovery now waits for durable settlement before continuing, shows recovery reasons and valid actions in WebShell, and gates message sending. (#15131) by @doudouOUC
• Embedding hosts can now show unavailable models as disabled rows in the composer picker, with hover or focus hints and optional host-provided actions. (#15015) by @callmeYe
• feat(managed-agent): Add session lifecycle UI and reading exports (#15100) by @doudouOUC
• feat(generated-ui): expose interaction outcomes and applied parameter snapshots (#15004) by @qqqys
• feat(managed-agent): expose durable capacity observations (#15162) by @yiliang114
• Pages Managed checkpoint history, adds durable model API error codes, and shows running indicators in Managed sidebar rows for clearer session feedback. (#15170) by @doudouOUC
• feat(core): gate scheduled tasks on a shell command run outside the model (#15069) by @wenshao
• Session lists now support case-insensitive title filtering before pagination, with archive filtering and a WebShell title search box. (#15157) by @yiliang114
• feat(managed-agent): read the Harness credentials from one file (#15063) by @wenshao
• feat(managed-agent): manage MCP servers without recreating sessions (#15119) by @doudouOUC
• feat(cli): align OpenTUI Vim, history and live output with Ink (#15067) by @chiga0
• feat(managed-agent): Enable public subagent delegation by default (#15171) by @doudouOUC
• feat(managed-agent): Start daemon and Managed locally with one command (#14928) by @doudouOUC
• feat(cli): ship private Bun and preserve interpreter boundaries (#14959) by @chiga0
• feat(mods): skip hooks for Claude Code events this runtime does not run (#15273) by @qqqys
• feat(mods): send a Mod's own tool.check ask to the user (#15306) by @qqqys
• feat(serve): carry a scheduled-task gate through the daemon and Web Shell (#15211) by @wenshao
• feat(ci): switchable Linux runner fleet (hosted/arm64) via LINUX_RUNNER_FLEET variable + dispatch option (#15329) by @yiliang114
• feat(serve): bind reviewed PR when a /review prompt is admitted (#15207) by @wenshao
• feat(cli): add a capability parity matrix against the Managed Agent path (#15085) by @wenshao
• feat(managed-agent): Enable unified Shell session lifecycle (#15330) by @doudouOUC
• feat(managed): gate managed automations on a Runtime probe run before the claim (#15218) by @wenshao
• feat(web-shell): let users reorder sidebar workspaces (#15360) by @wenshao
• feat(managed-agent): Add web_fetch to Workspace Sessions (#15313) by @doudouOUC
• feat(cli): add the PR watch probes behind the scheduled-task gate (#15057) by @qqqys
• feat(managed-agent): Add a shared Harness pool for Kubernetes (#15255) by @doudouOUC
• feat(verify-pr): configure the assets host and branch once (#15377) by @wenshao
• feat(web-shell): expose model output language in settings panel (#15363) by @wenshao
• feat(serve): give managed workspace sessions todo_write (#15076) (#15089) by @wenshao
• feat(serve): persist max-sessions in settings.json (#15419) by @wenshao
• feat(web-shell): add unified recent sessions sidebar (#14904) by @deggs7
• feat(mods): read session facts and the clock (#15348) by @qqqys
• feat(verify-pr): add a watch mode backed by the PR probe gate (#15409) by @wenshao
• feat(memory): add scoped external import jobs for CLI and HTTP (#14281) by @callmeYe
• feat(ci): fail fast when a pool host drifts off the font baseline (#15410) by @yiliang114
• feat(cli): attach a terminal to daemon sessions (#15209) by @wenshao
• feat(cli): add the /autofix-pr reaction layer: guarded push, sync, replies, re-runs and the skill (#15385) by @wenshao

Bug Fixes

• Release tarball checks now verify execute bits for every spawned vendored binary, including the Landlock runner, preventing silent permission regressions. (#15008) by @yiliang114
• Fixes an issue where directory changes could make a completed session appear interrupted, while keeping real pending prompts and tool calls recoverable. (#15012) by @ytahdn
• Fixes CI by enrolling OpenTUI parity integration tests in the no-AK lane so required test selection passes. (#15142) by @yiliang114
• Fixes a core TypeScript build failure by removing a duplicate import in a chat test, with no product behavior change. (#15146) by @ytahdn
• Gives fast Java CI jobs priority for SDK Java host-lock tickets while preventing starvation of full-tier jobs. (#15035) by @qwen-code-dev-bot
• Workflow agent subagent IDs now include the Workflow tool call ID when available, helping Web Shell link subagents to the call that started them. (#15145) by @qqqys
• Official DeepSeek support now uses deepseek-flash, accepts images for that exact model, and aligns effort mapping with DeepSeek's API. (#15164) by @qqqys
• The lint lane failure dump now reports swap alongside RAM and falls back to df -h when df -T is unavailable, improving resource diagnostics. (#15095) by @yiliang114
• The MariaDB lane now writes back only Maven artifacts fetched in that job and raises its timeout ceiling from 200 to 280 minutes. (#15064) by @yiliang114
• fix(web-shell): show actionable skill installation errors (#15013) by @callmeYe
• fix(web-shell): reduce table menu opening latency (#15179) by @ytahdn
• Restores a MariaDB integration leg in CI so the non-Hosted integration tests run against MariaDB 10.11.18 and MySQL 8.4. (#15003) by @yiliang114
• Aligns managed-checkpoint budgets, improves resource publication error details, and detects recovery-blocked Harness sessions even when heartbeats continue. (#15094) by @doudouOUC
• Updates Alibaba ModelStudio Token Plan models and thinking controls, adding deepseek-v4.1-flash variants and glm-5.3 while removing unavailable entries. (#15184) by @qqqys
• Refreshes the Alibaba ModelStudio Standard preset, enables model discovery, removes unavailable models, and adds deepseek-v4.1-flash and glm-5.3. (#15189) by @qqqys
• Anthropic global prompt-cache scope is now opt-in on every base URL, keeping per-session caching by default. (#15053) by @yiliang114
• Improves CI reliability by limiting the sdk-java host lock to Maven cache write-back and extending the integration timeout. (#15128) by @qwen-code-dev-bot
• fix(mods): exclude native waits from root execution budget (#14986) by @qqqys
• fix(core): cap MiniMax-M2.5 output at the DashScope endpoint limit (#15210) by @qqqys
• fix(serve): fall back to an available Local Control port (#12352) by @Sumire-no-kai
• fix(core): stop sed simulation from misreading backslashes in brackets (#13557) by @MohammedAlkindi
• test(managed-agent): poll ToolTurn producers instead of waiting on onExit() (#15065) by @wenshao
• fix(cli): recover from hosted send_message encoding limits (#14726) by @water-in-stone
• fix(web-shell): preserve composer draft attachments across views (#15025) by @ytahdn
• fix(cli): reject malformed hosted shell Unicode before admission (#13678) by @water-in-stone
• fix(web-shell): prevent artifact card overlap on narrow screens (#15300) by @ytahdn
• fix(managed-agent): refuse ; comments in the Harness credentials file (#15236) by @wenshao
• fix(managed-agent): drop removed publication constructor argument (#15314) by @qwen-code-dev-bot
• fix(ci): reuse open workflow failure issues (#15256) by @yiliang114
• fix(ci): gate the Windows Hosted suite on a successful install (#15290) by @qwen-code-dev-bot
• fix(web-shell): preserve draft attachments through refresh and allocation (#15303) by @ytahdn
• fix(core): bind session PRs when gh is outside PATH (#15212) by @wenshao
• fix(acp): stop blanking replay App html that cannot reclaim bytes (#15279) by @yiliang114
• fix(ci): refuse a partial jobs page list before reporting (#15174) by @wenshao
• fix(web-shell): honor loading phrase visibility setting (#15337) by @ytahdn
• fix(web-shell): Align Managed sidebar with daemon navigation (#15331) by @doudouOUC
• fix(mods): refuse values nested too deeply to cross the Mod boundary (#15317) by @qqqys
• fix(ci): install linters on linux/arm64 runners (#15350) by @wenshao
• fix(cli): kill a timed-out review command's process tree (#15352) by @wenshao
• fix(managed-agent): resume collection blockers at their own clearing instant (#15195) by @wenshao
• fix(cli): drop the removed V2 tool lists from the parity test (#15362) by @wenshao
• test(web-shell): cover reopened session details focus handoff (#15351) by @Jun516
• fix(release): skip cua-sdk publish dry-run when the version is already on npm (#15355) by @yiliang114
• fix(ci): annotate the temp short-alias guard's failing branches (#15366) by @yiliang114
• fix(ci): stabilize TUI attachment wait and allow full acceptance run (#15372) by @yiliang114
• test(core): give the paged-outcomes turn a CI-sized timeout (#15374) by @yiliang114
• fix(ci): route web shell smoke to ARM64 independently (#15364) by @yiliang114
• fix(node-repl): let the kernel exit promptly on shutdown (#15160) by @wenshao
• fix(ci): pin and cache the standalone runtime archives (#15159) by @wenshao
• fix(cli): install the reaping traps before the review command starts (#15367) by @wenshao
• fix(cli): keep reaping a review command's tree on a busy machine and under old bash (#15375) by @wenshao
• fix(ci): file new main CI failure signatures as their own issues (#15405) by @yiliang114
• fix(vscode): keep transcript selection alive during host-driven webview copy (#15403) by @yar3333
• fix(web-shell): move group creation into workspace menu (#15320) by @deggs7
• test(integration): give the skill-hook parity E2E the docker startup budget (#15226) (#15237) by @qwen-code-dev-bot
• fix(ci): raise the sdk-java test-lane lock wait above the longest hold (#15257) by @qwen-code-dev-bot
• fix(acp): refuse an edit whose arguments were cut off or arrived incomplete (#15346) by @qqqys
• fix(ci): skip stale main CI failure issues after main advances (#15272) by @qwen-code-dev-bot
• fix(cua): sign Windows UIAccess workers before publishing (#12442) by @yiliang114
• fix(cli): Replace synthesized tool-result placeholders with late real results (#15308) by @yiliang114
• fix(ci): skip the font baseline gate when the checkout predates it (#15436) by @yiliang114
• fix(cli): let review build-test reach the test phase on this monorepo (#15332) by @wenshao
• fix(core): count only well-formed snapshots toward the workflow retention cap (#15391) by @yiliang114

Performance

• Parallelizes MySQL and MariaDB integration tests, removes redundant validation work, and consolidates load scenarios to speed up CI. (#15200) by @doudouOUC
• perf(ci): scope PR lint and reuse static build outputs (#15381) by @yiliang114

Documentation

• Adds documentation for .agents/skills locations, the skills.directories setting, precedence, and reusing one skills folder with other agents. (#15086) by @shpaker
• Adds a bilingual design proposal for an opt-in Linux local-process confinement profile for Managed Agent, including qualification and fail-closed behavior. (#15163) by @yiliang114
• test(managed-agent): pin both worker launch paths to the env filter (#15051) by @wenshao
• test(managed-agent): align the FG12d reload comment with the lease expiry (#15234) by @wenshao
• docs(mods): propose the Claude Code compatibility baseline (#15270) by @qqqys
• docs(design): add the unified agent loop design (#15343) by @qqqys
• docs(verify-pr): add a --publish shorthand for maintainer rounds (#15357) by @wenshao
• docs(managed): document the PR watch probes as Managed automation gates (#15420) by @wenshao

Internal Changes

• Adds a test verifying that repeating a managed-agent cancel with the same idempotency key reuses the existing cancellation and starts no new work. (#14997) by @wenshao
• Internal test cleanup: shares the quote and camel helpers across managed-agent gate tests without changing behavior. (#15047) by @wenshao
• Updates Hosted fault gate tests to verify execution ownership, membership clearing, and scoped shutdown behavior. (#15066) by @doudouOUC
• Adds a CI guard ensuring every runnable integration test is selected by the no-AK gate or explicitly exempted, preventing accidental omissions. (#14989) by @zonemeen
• Improves internal CLI test reliability by resolving temporary Git fixture paths under symlinked directories. (#15133) by @wenshao
• test(managed-agent): refuse a second fire while the turn runs (FG11d) (#14998) by @wenshao
• Java SDK tests now mark body-less 404 responses as non-persistent so follow-up submissions use a fresh connection and avoid pooled connection failures. (#15172) by @qwen-code-dev-bot
• test(managed-agent): keep a sibling run alive past a cancel (FG11d) (#14999) by @wenshao
• test(core): pin that setTools writes the tool declarations (#15118) by @yiliang114
• test(managed-agent): pin durable recovery park acceptance (#15154) by @yiliang114
• Makes Hosted MCP fault test fixtures claim close operations atomically, preventing race conditions from interfering with intended fault scenarios. (#15158) by @yiliang114
• Adds a test ensuring the branch picker shows the Sets upstream on push hint when a branch has no upstream configured. (#15073) by @yiliang114
• Corrects the daemon E2E install warning and adds safeguards that keep Build, Bundle, and Java daemon E2E steps in the expected order. (#14884) by @qwen-code-dev-bot
• Adds a regression test ensuring the VS Code webview language attribute is HTML-escaped correctly. (#10492) by @CanReader
• Stabilizes the SDK Java idle-watchdog timing test without changing runtime watchdog behavior. (#15103) by @qwen-code-dev-bot
• test(managed-agent): keep the FG12d lifecycle cells' writer lease past the close window (#15068) by @wenshao
• test(managed-agent): pin the queue_one reachability (FG11d) (#15187) by @wenshao
• test(managed-agent): replace the Runtime MCP connection mid-stream (FG12e) (#15198) by @wenshao
• test(e2e): isolate permission-control suite from shared QWEN_HOME memory state (#11401) by @he-yufeng
• test(integration): give the resume gate E2E its CI startup budget (#15185) (#15188) by @qwen-code-dev-bot
• refactor(managed-agent): Remove pre-release compatibility paths (#15263) by @doudouOUC
• test(web-shell): pin which binding the session PR badge renders (#15238) by @wenshao
• test(managed-agent): fix ToolPublicationStore constructor call in HostedWorkspaceToolTurnIT (#15305) by @yiliang114
• test(managed-agent): share the Hosted gate attachment lookup (#15295) by @wenshao
• test(cli): cover workspace-scope aux model persistence in ModelDialog (#15356) by @yiliang114
• test(cli): stop assuming one build's CA loader and sleep semantics (#15370) by @yiliang114
• test(web-shell): pin running title cue spinner parity (#15371) by @Jun516
• test(web-shell): stop pinning hero header wrap in token usage min-width check (#15399) by @yiliang114
• test(web-shell): stop requiring the token-usage meta to wrap at min width (#15400) by @yiliang114
• test(web-shell): cross the hover-menu corridor with raw pointer moves (#15404) by @yiliang114
• ci: reduce redundant TUI parity runs and build scope (#15383) by @yiliang114
• refactor(web-shell): share the openable-PR filter with the details tooltip (#15373) by @wenshao
• test(scripts): pin newline-bearing takeover round seed bodies (#15390) by @yiliang114
• test(cli): widen held-expiry margins in the peer listing stale test (#15413) by @wenshao
• test(review): witness the unasserted halves of two Aone routing pins (#15395) by @yiliang114
• test(cli): pin the reaping shell's trap order and tighten its interrupt test (#15415) by @wenshao
• test(cli): poll the held-expiry assertions instead of sleeping (#15417) by @wenshao
• test(feishu): cover inbound file attachment lifecycle (#13278) by @jabrailkhalil
• test(managed-agent): retry the Hosted boot when a probed port is taken (#15432) by @yiliang114
• test(core): isolate outcomes-test retries from the dispatch gate (#15429) by @yiliang114
• test(runtime-broker): show what the rig saw when an execution wait times out (#15437) by @wenshao

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.26.0...v0.26.1

0.26.0 - 2026-10-21

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• feat(daemon): support session-targeted Skill and Extension refresh (#14050) by @callmeYe
• feat(runtime-broker): Session-create warm for Workspace-isolated bound Sessions (#14599) by @doudouOUC
• feat(managed-agent): ship a configuration template in the reference distribution (#14631) by @wenshao
• feat(serve): run private CSI file turns through Hosted Harness (#14487) by @doudouOUC
• feat(managed-agent): support local and SSH workspace file references (#14639) by @doudouOUC
• Adds a read-only --check preflight that prints PASS, WARN, FAIL, or SKIP for managed-agent deployment checks and exits nonzero on failure. (#14632) by @wenshao
• Lets a successor Hosted owner finish private CSI file turns that already have durable receipts, preserving tool responses and avoiding repeated file operations. (#14523) by @doudouOUC
• Managed Agent submissions now accept up to 8 MiB of attachment input, preserve large pending requests across refresh, and let users discard unrecoverable drafts. (#14628) by @doudouOUC
• Hosted Turn recovery now asks the Runtime Broker for execution evidence and treats unprovable outcomes as unknown instead of transient faults. (#14669) by @wenshao
• Adds systemd units and setup guidance for single-host managed-agent deployments so Runtime workers stay running when Broker restarts (#14633) by @wenshao
• Parks managed-agent turns with unknown Runtime outcomes for later rechecks instead of repeatedly retrying cancellation, keeping them active and avoiding false FAILED results (#14681) by @wenshao
• feat(artifacts): record files the agent creates or edits (#14292) by @zjgzx1988
• Adds a managed-agent credential bootstrap that generates required deployment tokens and writes protected configuration to simplify secure setup. (#14634) by @wenshao
• Completes private CSI cold recovery so settled file work can resume from saved results without repeating file effects or extra model requests. (#14695) by @doudouOUC
• Adds an optional Generated UI extension that renders model-generated HTML/CSS cards with bounded local JavaScript, letting users adjust inputs without another model turn. (#14338) by @qqqys
• feat(workflow): authorize daemon ancestor workflows (#14646) by @qqqys
• feat(web-shell): show per-model TTFT and output speed in Token Usage (#14728) by @qqqys
• feat(runtime): add original CSI application retirement (#14698) by @doudouOUC
• feat(telemetry): Record reasoning effort on LLM spans (#14718) by @zonemeen
• feat(serve): add local daemon discovery and control commands (#14739) by @wenshao
• feat(web-shell): Polish advanced table controls and copying (#14707) by @ytahdn
• feat(core): account tool result sizes from producer to injection (#14545) by @yiliang114
• feat(web-shell): add trajectory follow and pause controls (PR7B) (#14723) by @qqqys
• feat(web-shell): delete selected message groups from session history (#14297) by @ytahdn
• feat(core): honor preview budgets and make truncated previews recoverable (#14546) by @yiliang114
• feat(web-shell): Preview collapsed Home content on hover (#14826) by @ytahdn
• feat(web-shell): preview PDFs with the native browser viewer (#14825) by @ytahdn
• feat(web-shell): improve token usage hierarchy (#14648) by @ytahdn
• feat(mods): support tool middleware in interactive terminals (#14565) by @qqqys
• feat(web-shell): show background shell task runtime details (#14831) by @ytahdn
• feat(core): report why loop protection stopped a turn (#14858) by @qqqys
• feat(serve): add qwen daemon start (#14869) by @wenshao
• feat(channels): add WeCom application callback support (#13976) by @callmeYe
• feat(serve): add daemon status detail (#14902) by @wenshao
• feat(cli): expose experimental Mods in settings (#14889) by @qqqys
• feat(web-shell): preview DOCX files across file entry points (#14868) by @ytahdn
• feat(web-shell): refresh model catalogs from the model selector (#14685) by @callmeYe
• feat(serve): add daemon start --wait (#14914) by @wenshao
• feat(managed-agent): expose session MCP catalog in WebShell (#14916) by @doudouOUC
• feat(runtime): add explicit idle resident Workspace retirement (#14765) by @doudouOUC
• feat(web-shell): consolidate trajectory chrome into a summary bar and toolbar (#14913) by @qqqys
• feat(managed-agent): Support approval changes during a turn (#14920) by @doudouOUC
• feat(managed-agent): simplify existing local project onboarding (#14970) by @doudouOUC
• feat(generated-ui): bundle Generated UI behind an experimental flag (#14888) by @qqqys
• feat(web-shell): show model refresh as an icon beside search (#14983) by @callmeYe
• feat(managed-agent): import enabled skills at session creation (#14930) by @doudouOUC
• feat(serve): add daemon restart (#14929) by @wenshao
• feat(cua): streamline computer use and port native reliability fixes (#14846) by @LaZzyMan
• feat(web-shell): support managed session rename (#14987) by @doudouOUC
• feat(serve): add persisted workspace artifact library (#14867) by @BZ-D
• feat(workflow): add per-turn human-origin workflow hints (#14876) by @qqqys
• feat(web-shell): open trajectory record details as a drill-in or side panel (#14984) by @qqqys
• feat(core): add a real PTY backend for the Bun runtime (#14046) by @chiga0
• feat(managed-agent): allow concurrent sessions on shared workspace storage (#15026) by @doudouOUC
• feat(managed-agent): derive Session Store wiring on a single host (#15005) by @wenshao
• feat(serve): make paired Web Shell devices listable and revocable (#15000) by @yiliang114
• feat(core): add model.loopProtection and remind instead of stopping loops by default (#15016) by @qqqys
• feat(cli): paste clipboard text with Ctrl+V on Linux (#14172) by @qwen-code-dev-bot
• feat(web-shell): unify daemon and managed workspace navigation (#15061) by @doudouOUC

Bug Fixes

• fix(managed-agent): bound initial publication guard waits (#14563) by @yiliang114
• test(cli): retry ENOTEMPTY in hosted harness teardowns (#14636) by @ZedingZhang
• Automatically recovers public Sessions with committed, unconsumed inputs after managed-agent writer loss so pending wakes are replayed instead of dropped. (#14603) by @doudouOUC
• Session creation and worktree reset via the daemon HTTP API now use the daemon's session budget deadline, preventing in-progress initialization from timing out too early. (#14588) by @yiliang114
• On POSIX, Hosted stdio MCP servers now run in their own process group with supervised shutdown, improving cleanup; Windows behavior is unchanged. (#14640) by @wenshao
• Hosted Session journals now recover after a lapsed writer lease, applying committed entries once and retrying missing ones instead of failing. (#14570) by @wenshao
• Hosted PostToolUse, PostToolUseFailure, and PostToolBatch hooks now report the tool input left by PreToolUse, excluding refused or invalid replacements. (#14650) by @qqqys
• fix(ci): repair Java gates and report failures on PRs (#14522) by @yiliang114
• Hosted sessions now continue recovered child-result and monitor notification turns after a crash instead of waiting for new input. (#14679) by @doudouOUC
• Raises Playwright wait limits for web-shell smoke tests from 10 to 20 seconds to reduce false failures on busy shared CI hosts (#14643) by @yiliang114
• Fixes managed-agent tool cards so expanding a settled ReadFile or Shell card loads saved arguments, output, and errors (#14604) by @doudouOUC
• Preserves keyboard focus in the web-shell trajectory panel when recent execution views are removed or evicted (#14641) by @qqqys
• fix(cli): reconcile managed recovery failure reasons (#14567) by @yiliang114
• fix(hooks): reap a Hook command unit when its worker dies (#14502) (#14675) by @wenshao
• fix(config): let a Managed host restore its own session (#14688) by @wenshao
• Widens interactive E2E wait budgets for saved workflow slash-command completion replies to 120 seconds to reduce flaky CI stalls. (#14202) by @qwen-code-dev-bot
• Prevents a pending user cancellation from being recorded as a failed turn when recovery is not ready, allowing it to remain cancelling or settle as cancelled. (#14694) by @wenshao
• Bounds managed-agent publication placement waits to prevent unlimited blocking and long lock timeouts during dispatch, completion, claim, and verification. (#14542) by @yiliang114
• Managed Agent image attachments now accept PNG, JPEG, GIF, WebP, and the image/jpg alias, matching Qwen Code image handling. (#14704) by @doudouOUC
• Stabilizes the FG12b released-store fault gate so it reliably verifies owner-release resend behavior without depending on CI load. (#14711) by @qwen-code-dev-bot
• Fixes CI so Java failure summaries can be posted as pull request comments by granting the comment writer job pull-request write permission. (#14717) by @yiliang114
• fix(managed-agent): reject an incomplete Workspace tool path in --check (#14635) by @wenshao
• fix(cli): hold pre-runtime /acp WebSocket upgrades instead of falling through to a 406 (#13995) (#13997) by @qwen-code-dev-bot
• fix: align DeepSeek reasoning controls across official and MaaS routes (#14689) by @qqqys
• fix(web-shell): surface the fix witness in the review artifact viewer (#14725) by @yiliang114
• fix(managed-agent): align fault gates with unsettled-append recovery (#14716) by @yiliang114
• test(managed-agent): widen the hosted recovery and hook-cancel gate windows (#14729) (#14741) by @qwen-code-dev-bot
• fix(web-shell): preserve slash highlighting in user messages (#14171) by @Lilian0122
• fix(managed-agent): prove a zombie-only group stopped (#14738) by @wenshao
• test(managed-agent): drive the FG12b releasing gate until the fault fires (#14782) (#14799) by @qwen-code-dev-bot
• test(managed-agent): pin the FG13c worker-kill reaper contract (#14802) by @qwen-code-dev-bot
• test(managed-agent): give the failed-spawn test a realistic deadline (#14779) (#14783) by @qwen-code-dev-bot
• fix(core): Compact managed activation and inbox journals (#14760) by @doudouOUC
• fix(managed-agent): Reconcile parked Turn cancels from durable results (#14763) by @doudouOUC
• fix(ci): repair main's Prettier step and deflake the hook-cancel fence test (#14581) (#14590) by @qwen-code-dev-bot
• fix(ci): wait 90 minutes for the sdk-java host lock (#14787) (#14801) by @qwen-code-dev-bot
• perf(core): Index managed Session history reads (#14767) by @doudouOUC
• fix(managed-agent): park a cancel over an unobserved Turn (D7-2) (#14700) by @wenshao
• fix(ci): include updatedInput hook integration tests in PR checks (#14747) by @yiliang114
• fix(ci): allow fault gates to outlive the Surefire fork timeout (#14745) by @yiliang114
• fix(cli): keep one-shot system reminders out of the user's own message (#14740) by @yiliang114
• fix(ci): raise the build heap cap from 4096 to 6144 MB (#14822) by @yiliang114
• fix(acp): avoid splitting surrogate pairs when truncating turn text (#14744) by @yiliang114
• fix(managed-agent): Preserve original owners for owed Hosted cleanup (#14764) by @doudouOUC
• fix(core): keep a producer-declared truncation bit when raw size is unknown (#14830) by @yiliang114
• fix(ci): cap the prepare build heap and share the self-hosted Node preflight (#14827) by @wenshao
• fix(ci): skip failure reports for superseded run attempts (#14768) (#14771) by @qwen-code-dev-bot
• fix(managed-agent): answer self-healing Harness conflicts as retryable 503 (#14645) by @wenshao
• fix(web-shell): avoid flashing the credential form during boot token verification (#14770) by @BigDaddy1337
• fix(core): stop double-truncating web_fetch and keep the full text recoverable (#14547) by @yiliang114
• fix(managed-agent): keep the CSI retirement key mode-independent on MariaDB (#14832) (#14853) by @yiliang114
• fix(hosted): bind execution to the final hook input (#14708) by @qqqys
• test(managed-agent): stabilize the FG12b intent and cwd hosted gates (#14730) by @qwen-code-dev-bot
• fix(managed-agent): treat hosted_turn_active as a lifecycle wait (#14492) (#14806) by @qwen-code-dev-bot
• fix(ci): retry transient pnpm installs in the SDK Java workflow (#14845) (#14851) by @qwen-code-dev-bot
• fix(channels): preserve Markdown structure when splitting messages (#14043) by @callmeYe
• fix(memory): report the real cause and drop set when MEMORY.md exceeds its budget (#14841) by @yiliang114
• fix(ci): skip Managed Agent reviews after the PR closes (#14774) (#14777) by @qwen-code-dev-bot
• fix(managed-agent): bound settled runtime dispatch entries (#14657) by @wenshao
• fix(ci): skip prepare's duplicate build in the daemon E2E install (#14861) (#14873) by @qwen-code-dev-bot
• fix(cli): keep workspace skill status aligned after session load (#14840) by @callmeYe
• test(managed-agent): Measure the post-commit gate as per-round growth (#14197) (#14207) by @qwen-code-dev-bot
• fix(managed-agent): refresh schema upgrade expectations for V12/V13 (#14893) by @qwen-code-dev-bot
• fix(managed-agent): bound and harden session resource storage (#14656) by @wenshao
• fix(managed-agent): stop activation renewal after persistent failure (#14658) by @wenshao
• test(cli): give resident model recovery waits the slow-host budget (#14922) (#14927) by @qwen-code-dev-bot
• fix(ci): hold the MariaDB lane's sdk-java lock once per job (#14896) (#14903) by @qwen-code-dev-bot
• fix(ci): hold chosen RocketMQ ports until the JVMs bind them (#14860) by @yiliang114
• fix(release): publish the generated CLI with npm so vendored ripgrep keeps its exec bit (#14907) by @yiliang114
• test(managed-agent): run the MariaDB lane's fixtures on MariaDB (#14906) by @yiliang114
• test(ci): pin the heap caps that failed the Managed Agent fast tier (#14862) (#14874) by @qwen-code-dev-bot
• test(integration): wait for the CLI PTY to exit before resetting a case directory (#14957) (#14960) by @qwen-code-dev-bot
• fix(ci): isolate hosted MySQL integration test classes (#14875) by @doudouOUC
• test(managed-agent): widen the held-approval lease margins against CI scheduler jitter (#14478) by @qwen-code-dev-bot
• fix(core): preserve compressed context during recovery (#14702) by @ytahdn
• test(integration): give each workflow completion case its own test directory (#14792) by @qwen-code-dev-bot
• fix(cli): restore bounded OpenTUI dialogs and add migration gates (#14933) by @chiga0
• fix(test): keep E2E CI runs off the pool host's auth state (#14972) by @qwen-code-dev-bot
• fix(web-shell): align Managed Session titles, file access and the Workspace row (#14924) by @yiliang114
• fix(ci): queue the sdk-java host lock first-come-first-serve (#14743) (#14912) by @yiliang114
• fix(cli): preserve runtime boundaries during standalone updates (#14934) by @chiga0
• fix(cua): restore release version validation from README (#15001) by @LaZzyMan
• fix(cli): keep review's gh reads machine-readable under forced colour (#14944) by @yiliang114
• fix(cli): report a zenity 4 display failure instead of a silent cancel (#14923) by @yiliang114
• fix(serve): admit loopback origins on any port via --allow-origin (#15002) by @yiliang114
• fix(core): preserve Responses cache boundaries before auto-memory context (#15006) by @LaZzyMan
• fix(web-shell): preserve reading position across SSE reconnects (#15009) by @ytahdn
• fix(managed-agent): keep control-plane settings out of Runtime workers (#15024) by @wenshao
• fix(core): place DashScope conversation cache before the attachment-bearing message (#15020) by @yiliang114
• fix(core): keep a close admitted mid-request from wedging the Hosted writer (#15010) by @wenshao
• fix(web-shell): keep the sidebar PR badge still when row actions appear (#14994) by @wenshao
• test(cli): keep daemon workflow intent E2E off the container sandbox (#15029) by @qwen-code-dev-bot
• fix(core): pluralize the Goal token budget stop reason (#15062) by @yiliang114
• fix(ci): align Java integration database and lock lifetime (#14936) by @yiliang114
• fix(ci): name the sdk-java host lock holder when a ticket starves (#15031) by @yiliang114

Performance

• Shell tool receipt lookups are faster because committed receipts are indexed by execution call instead of scanning full history. (#14660) by @wenshao
• Improves web-shell performance by incrementally updating managed session messages when new events append, avoiding full history rebuilds (#14562) by @yiliang114
• perf(managed-agent): bound the held-Runtime inventory read cost (#13930) (#14691) by @wenshao

Documentation

• Adds a bilingual design document proposing bounded retention, replay floors, snapshots, and windowed event storage for managed-agent session state. (#14659) by @wenshao
• Clarifies VS Code documentation on why embedded webviews cannot use the standalone Web Shell same-origin exception for daemon URLs. (#14684) by @Jun516
• test(managed-agent): finish the FG13f decision-delivery arms (#14736) by @wenshao
• docs(managed-agent): name the pre-dispatch commit-spy assertion (#14756) by @wenshao
• docs(managed-agent): record the D7-2 report as closed out (#14800) by @wenshao
• docs(managed-agent): scope the FG11f publication fault gates (#14734) by @wenshao
• docs(managed-agent): pin the reconcile-before-refusal and the OBJECT_PUT admission (#14811) by @wenshao
• docs(managed-agent): record the FG13f remaining-arm coverage (#14814) by @wenshao
• docs(managed-agent): close out the D7 report statuses (#14813) by @wenshao
• docs(runtime): simplify CSI delivery around resident Workspaces (#14758) by @doudouOUC
• docs(telemetry): clarify OTLP HTTP uses JSON encoding (#14664) by @Jun516
• docs(managed-agent): scope the FG11e stream-gap gate (#14881) by @wenshao
• docs(managed-agent): record what the remaining FG11d cells answer (#14882) by @wenshao
• docs(managed-agent): locate the FG11e replay branch and gap contract (#14899) by @wenshao
• docs(managed-agent): name the replay contract document (#14905) by @wenshao
• docs(mcp): clarify restrictive prefixes and legacy grants (#14683) by @Jun516
• docs(managed-agent): propose where a late reply settles after the connection is gone (#14926) by @wenshao
• docs(board): clarify suggested task owner semantics (#14967) by @Jun516
• docs(ci): correct stale MariaDB CI claims and host reaper bound (#14995) by @yiliang114

Internal Changes

• test(cli): give turn-settling waits the slow-host budget (#14598) by @yiliang114
• Raises CLI hosted probe-start test guards from 1 second to 10 seconds to tolerate scheduling delays and reduce false failures. (#14662) by @yiliang114
• Adds a managed-agent test verifying DELETE drains a held async PreToolUse hook before SessionEnd and SessionDelete lifecycle hooks run. (#14638) by @wenshao
• Internal tests now verify automation recovery after Spring crashes and scanner lease races. (#14649) by @wenshao
• Internal cleanup that shares CSI runtime marker reads without changing managed-agent behavior (#14571) by @yiliang114
• Adds a managed-agent fault test ensuring a lost hook-catalog reply before registration does not create duplicate or stale Hook registrations (#14668) by @wenshao
• test(managed-agent): assert G3 approval identities and pin drift (#14673) by @wenshao
• test(managed-agent): drop the FG13a HTTP Hook reply after the request (#14692) by @wenshao
• test(managed-agent): scope the FG11c crash assertions and fix the race fixture (#14682) by @wenshao
• Adds tests confirming the Session-close drain flag is sent and handled correctly during managed-agent release requests. (#14712) by @wenshao
• Adds a regression test ensuring session labels are recovered only from verified Goal records, not lines that merely mention the marker. (#14705) by @yiliang114
• Removes the fixed JavaScript size ceiling for read-only transcript artifact checks to reduce recurring test maintenance and merge conflicts. (#14703) by @ytahdn
• Reorganizes Managed Agent Server internal packages by responsibility without changing deployment or service behavior. (#14710) by @doudouOUC
• Adds G3 approval fault-point end-to-end arms that verify durable approval recovery after crashes, lost replies, and dispatch failures. (#14687) by @wenshao
• Synchronizes scheduler disposal fault-injection testing to reliably count failed claims before disposal without fake-clock interference. (#14727) by @yiliang114
• test(managed-agent): replay a denied Hook decision after a lost reply (#14701) by @wenshao
• test(cli): pin the FG13d pre-dispatch cancel in a unit test (#14735) by @wenshao
• test(managed-agent): witness the durable hold sweep across both families (#13930) (#14719) by @wenshao
• test(managed-agent): gate a disabled automation span (FG11d) (#14731) by @wenshao
• test(managed-agent): leave a v2 execution unknown across a Harness kill (D7-2) (#14750) by @wenshao
• test(managed-agent): cover the remaining FG12a reply-loss cells (#14715) by @wenshao
• test(managed-agent): replay a failClosed Hook failure after a lost reply (#14803) by @wenshao
• test(managed-agent): Verify private media retention boundaries (#14766) by @doudouOUC
• refactor(managed-agent): simplify the event transport package (#14816) by @doudouOUC
• test(managed-agent): refuse the first FG12d release before forwarding (#14737) by @wenshao
• test(ci): pin that the temp short-alias guard visits both keys (#14805) by @yiliang114
• test(managed-agent): Verify MCP retention through Session retirement (#14762) by @doudouOUC
• test(managed-agent): let the FG13f failClosed cell reach the Hook (#14818) by @wenshao
• test(managed-agent): settle a landed turn reply without a cold load (#14820) by @yiliang114
• test(managed-agent): crash the approval result-commit barrier (#14733) by @wenshao
• test(core): pin the durability snapshot after reactive compaction (#14815) by @wenshao
• test(managed-agent): reject Hook store commits (FG13b) (#14817) by @wenshao
• test(managed-agent): gate the FG11f lost-ack publication fault (FG11f) (#14819) by @wenshao
• test(managed-agent): gate FG12f result and catalog delivery (#14829) by @wenshao
• test(ci): pin the root build heap cap above the CLI compile ceiling (#14835) (#14843) by @qwen-code-dev-bot
• test(web-shell): stabilize branch picker workspace-switch focus test (#14842) by @qwen-code-dev-bot
• ci(gate): skip the Managed Agent gate run for reviews that are not approvals (#14857) by @yiliang114
• test(managed-agent): gate the FG11f refusing arms (FG11f) (#14821) by @wenshao
• test(managed-agent): wait for the landed reply's applied record (#14833) by @yiliang114
• test(managed-agent): ignore transient MySQL deadlocks in hosted gate waits (#14812) by @qwen-code-dev-bot
• style(managed-agent): clear the checkstyle violations in the result-delivery IT (#14883) by @wenshao
• test(managed-agent): reject Hook result and plan-result commits (FG13b) (#14848) by @wenshao
• test(managed-agent): pin which commit the plan rejection hit (#14898) by @wenshao
• test(managed-agent): gate the per-run cancel on a parked wake turn (FG11d) (#14900) by @wenshao
• test(integration): stop the session-id uniqueness E2E at the init message (#14579) (#14585) by @qwen-code-dev-bot
• test(managed-agent): gate a late conclusive reply after outcome_unknown (#14892) by @wenshao
• test(managed-agent): settle a claim retired at intent (FG11d) (#14915) by @wenshao
• test(managed-agent): skip a closed Session's automation slots (FG11d) (#14931) by @wenshao
• test(managed-agent): resume both Session streams across a held Hook (FG13e) (#14917) by @wenshao
• test(managed-agent): pin the close refusal while a turn is active (FG11d) (#14974) by @wenshao
• test(managed-agent): gate a pre-dispatch MCP operation cancel (FG12d) (#14963) by @wenshao
• test(managed-agent): keep a suppressed reply suppressed across a reload (FG13f) (#14978) by @wenshao
• test(managed-agent): reuse the run on a same-key retry (FG11d) (#14981) by @wenshao
• test(managed-agent): gate R1-1 owner-release recovery across a Spring restart (FG12d) (#14990) by @wenshao
• test(managed-agent): keep a settled run's outcome (FG11d) (#14996) by @wenshao
• test(managed-agent): lose the store ack after the result commit (FG13a) (#14991) by @wenshao
• test(managed-agent): keep the durable dispatch after a lost ack (FG13a) (#14992) by @wenshao
• test(managed-agent): reuse the pinned registration after a lost ack (FG13a) (#14993) by @wenshao
• test(managed-agent): resume the automation streams across a settle (FG11e) (#15017) by @wenshao
• test(web-shell): pin the openable session PR filter directly (#15059) by @wenshao
• test(managed-agent): wait for the D7-2 runtime park, not a takeover refusal (#15046) by @yiliang114

New Contributors

• @Jun516 made their first contribution in #14684
• @BigDaddy1337 made their first contribution in #14770

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.25.3...v0.26.0

0.25.3 - 2026-10-19

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• Adds optional model-generated layouts to Test Failure Explorer, with validation and fallback to the fixed view (#14269) by @qqqys
• The Web Shell mobile composer no longer shows the ↑ and ↓ history buttons, reducing clutter while keeping history search available. (#14220) by @pomelo-nwu
• The Web Shell trajectory panel now includes a collapsible Recent executions section that shows live requests and tools with elapsed time and final durations. (#14071) by @qqqys
• Managed sessions created from published AgentDefinitions can now load static Skills and pass pinned MCP server and Hook catalog selections to Hosted Runtimes. (#14319) by @doudouOUC
• Files now offers New file and New folder actions, and folder Delete asks for a second confirmation before permanently removing the folder and its contents. (#14169) by @ytahdn
• Eligible legacy cancellations now require confirmation through the daemon or WebShell Continue flow before continuing, while explicitly user-cancelled work stays blocked. (#14087) by @yiliang114
• Hosted parent agents can launch snapshot-isolated child agents that read a frozen workspace copy without changing files, running commands, or affecting the parent. (#14265) by @wenshao
• Clicking the workspace folder beside the chat title now opens and selects the Files tree in the right panel, with keyboard and focus preserved. (#14344) by @ytahdn
• Managed sessions now integrate into the home and sidebar, supporting workspace creation, approval policy and model selection, and context usage in the composer. (#14316) by @doudouOUC
• The Files panel adds a Delete menu for regular files, requiring a permanent-deletion confirmation for the selected trusted local workspace. (#14158) by @ytahdn
• Adds resident Kubernetes Workspace file sessions that retain an original Pod and fixed CSI PVC so configured Workspaces preserve files across sessions. (#14340) by @doudouOUC
• Adds Shell Full Access and default/auto-edit/yolo switching for built-in Workspace-bound qwen-code Sessions without losing history, Workspace, or drafts. (#14323) by @doudouOUC
• feat(mods): support commands in interactive terminals (PR4a) (#14268) by @qqqys
• feat(providers): add QwenCloud as the first /auth entry (#14073) by @yiliang114
• feat(ci): route the release lane's hosted stragglers onto ECS runners (#14336) by @yiliang114
• feat(web-shell): add side task archive controls (#14300) by @yiliang114
• feat(managed-agent): Add SSH Tool Runtime support (#14318) by @doudouOUC
• feat(web-shell): list every Managed task kind in the task panel (#14415) by @wenshao
• feat(managed-agent): H4f-c public task cancel for background Shell and Monitor tasks (#14412) by @wenshao
• feat(managed-agent): link a child task to its child Session (#13746 F2) (#14413) by @wenshao
• feat(managed-agent): Improve workspace selection and tool progress (#14405) by @doudouOUC
• feat(managed-agent): project channel deliveries as tasks and serve their cancel (#13746 F3) (#14414) by @wenshao
• feat(managed-agent): record Turn input and cancel as durable command operations (D7-1) (#14404) by @wenshao
• feat(managed-agent): support native macOS workspace runtime (#14442) by @doudouOUC
• feat(qqbot): accept inbound files and voice messages (#14343) by @Eric-GoodBoy-Tech
• feat(managed-agent): reconcile acceptance-unknown operations through one rule (D7-2) (#14443) by @wenshao
• feat(managed-agent): reconcile a parked SSH Workspace registration (D7-2) (#14456) by @wenshao
• feat(runtime-broker): Workspace-resident Tool Runtime for bound Sessions (#14435) by @doudouOUC
• fix(ci): run three hosted control-plane jobs on the ECS pool (#14469) by @wenshao
• feat(managed-agent): add the opt-in java_durable admission profile (D7-3) (#14470) by @wenshao
• feat(ci): gate Managed Agent changes on a full SDK Java run before merge (#14397) by @yiliang114
• feat(serve): attach private CSI Sessions to Hosted text turns (#14081) by @doudouOUC
• feat(managed-agent): reconcile parked Session title deliveries (D7-2) (#14475) by @wenshao
• feat(managed-agent): register the Managed engine in paired daemons (#14483) by @wenshao
• feat(managed-agent): allow approval changes in stored agent sessions (#14484) by @doudouOUC
• feat(managed-agent): support Plan mode and inline attachments (#14531) by @doudouOUC
• feat(workflow): discover trusted ancestor workflows in local sessions (#14369) by @qqqys
• feat(managed-agent): Support per-run automation cancellation (#14526) by @doudouOUC
• feat(managed-agent): bound half-open Hosted event streams behind a keepalive capability (#14135) (#14512) by @wenshao
• feat(managed-agent): connect the composer to pinned Skills (#14602) by @doudouOUC
• feat(web-shell): add approval mode allowlist (#14345) by @ytahdn

Bug Fixes

• Fixes plan approval crashes during journal-only interrupted turns so takeover can settle the session instead of failing terminally (#14160) by @wenshao
• Fixes checkstyle build failures in the Java SDK by reordering FG8 test imports (#14305) by @wenshao
• Standardizes Managed Agent timestamps to Unix epoch milliseconds and updates query mappings for consistent data handling (#14185) by @doudouOUC
• Fixes CSI preparation failures by retaining input, tool declaration, and intent resources through the HTTP store (#14180) by @doudouOUC
• Managed Hosted model prompts now stay aligned with declared tools, including empty or changed tool sets, so tool guidance and examples remain accurate. (#14306) by @doudouOUC
• Closing a parent session now stops each child session's running Turn, reducing lingering background work and making close behavior more reliable. (#14218) by @wenshao
• Session close after a Harness restart or takeover now correctly tracks and releases held background Shell or Monitor runtimes, preventing abandoned runs. (#13986) by @wenshao
• fix(managed-agent): name the plan's size and room in the team-plan bound refusal (#14057) by @wenshao
• Windows merge checks now report known workspace and script failures as warnings while keeping installer, security, and Hosted process tests blocking. (#14298) by @yiliang114
• Managed-agent retry loops now have budgets and terminal states, letting stuck projections fail safely, reducing repeated log noise and preventing endless retries. (#13219) by @wenshao
• When a hosted member message reply times out, the CLI now advises not to resend because the recipient may already have it or receive it later. (#14296) by @yiliang114
• Dialogs opened over the mobile sidebar drawer remain touch-scrollable instead of being locked by the drawer's background scroll behavior. (#14217) by @pomelo-nwu
• Managed agent publication grants and background session close now work reliably for Shell and Monitor output. (#14364) by @doudouOUC
• Fixes CLI custom commands so injected file content is preserved during later shell and argument interpolation instead of being reinterpreted. (#14295) by @yiliang114
• Fixes Web Shell so Show Response Tokens Per Second displays an estimated t/s rate beside the streaming token counter when enabled. (#14294) by @yiliang114
• Polishes Web Shell file display and preview by showing safe file names, preserving paths, adding a close-preview action, and limiting syntax highlighting resources. (#14143) by @ytahdn
• Routes post-merge classifier jobs to ECS when enabled and lets CLI test workers yield between tests to drain queued progress replies. (#14293) by @yiliang114
• fix(hooks): gate headless file and skill prompts before expansion (#14356) by @qqqys
• fix(ci): make finalize-release second-bot approve non-blocking (#14341) (#14342) by @yiliang114
• fix(managed-agent): hold behind an older message whose body cannot be read (#14299) by @wenshao
• fix(managed-agent): tenant-wide InnoDB locks must not wedge unrelated Sessions (#13918) by @wenshao
• fix(managed-runtime): landed PR 12691 review follow-ups across providers, activator, core tools and runtime broker (#13297) by @wenshao
• fix(channels/email): re-register on a channel_disconnected answer (#13768) by @wenshao
• fix(sdk-java): Close Hosted Harness review criticals from #12654 (#13314) by @wenshao
• fix(managed-agent): hand back the Workspace mount when a tool turn goes recovery blocked (#13816) by @wenshao
• fix(managed-agent): Report merges deleting uncommitted parent content as conflicts (#14058) by @wenshao
• fix(ci): document the shared-group supersede policy and keep main uncancellable (#14286) by @yiliang114
• fix(serve): fan out never-started process probes (#14303) by @yiliang114
• fix(web-shell): inset the right edge of a label-only composer chip (#13968) by @Bug-killer-prog
• fix(web-shell): neutralise controls in the generic approval content block (#13972) by @Bug-killer-prog
• fix(web-shell): delete attachments from an empty composer (#13947) by @Zhangchenhao-111
• fix(managed-agent): close the deferred #13182 terminal-state suggestions (#14358) by @wenshao
• fix(managed-agent): log the check behind a refused internal request (#14391) by @wenshao
• fix(ci): assemble the reference distribution into a fresh directory (#14421) by @wenshao
• fix(daemon): preserve active work when deleting scheduled tasks (#14276) by @yiliang114
• fix(core): report a defunct session pid as dead instead of active (#14301) by @yiliang114
• fix(runtime-broker): support credential key rotation (#14317) by @yiliang114
• fix(ci): keep the setup-java Maven cache on hosted runners (#14396) by @yiliang114
• fix(managed-agent): run a merge recorded on a snapshot as a discard (#13905 items 33-36) (#14432) by @wenshao
• fix(ci): start the reference distribution MySQL without native AIO (#14429) by @yiliang114
• fix(web-shell): Distinguish unavailable shell output (#14436) by @doudouOUC
• fix(ci): stop re-running the H2 unit suite in every Hosted job step (#14423) by @yiliang114
• fix(managed-agent): address the H4f task cancel review follow-ups (#14437) by @wenshao
• fix(managed-agent): flush early reject responses before draining the unread body (#14444) by @wenshao
• fix(web-shell): improve favicon contrast on dark browser tabs (#14417) by @wenshao
• fix(managed-agent): answer a full tool publication reservation with its own code (#14451) by @wenshao
• fix(serve): answer a batch the publication limit refused instead of blocking the Session (#14452) by @wenshao
• fix(managed): Recover cancellation after process restarts (#14377) by @doudouOUC
• fix(managed-agent): skip the deletion probe when a resume owes nothing (#13905 items 50-52) (#14462) by @wenshao
• fix(vscode): avoid navigator reads in the extension host (#14304) by @zonemeen
• fix(managed-agent): address the record/contract review follow-ups of the H3 enablement (#13930) (#14464) by @wenshao
• test: fix the SDK Java suite's red jobs (FG6e frames, schema V8 pin, client flakes) (#14457) by @yiliang114
• fix(ci): give the hosted Test lane the same 120-minute ceiling as ECS (#14474) by @wenshao
• fix(ci): run the trusted web-shell browser smoke on the ECS pool (#14471) by @wenshao
• fix(ci): skip application tests for Markdown-only and CODEOWNERS-only changes (#14284) by @yiliang114
• fix(core): fill empty signature on unsigned thinking blocks for strict Anthropic-compatible proxies (#13676) by @4ekuct25
• fix(managed-agent): never re-drive an unsettled workflow Turn on the recovery path (#14387) (#14461) by @wenshao
• test(core): size the hosted cold-scan ceiling from the slowest hosted runners (#14497) by @wenshao
• test(managed-agent): stub the acquire provisioner overload the Broker calls (#14498) by @wenshao
• test(managed-agent): pin #14387's Turn-row accessors in the lock canary (#14501) by @wenshao
• fix(managed-agent): keep a crash-blocked workflow run unknown and honestly notified, gated by FG10d (#14165) (#14506) by @wenshao
• test(managed-agent): parse only complete rows in the wedge driver wire (#14496) by @wenshao
• fix(managed-agent): keep runtime MCP state and permission sync honest (#14488) by @wenshao
• fix(sdk-java): keep the snapshot flag when a load copy restates child Workspaces (#13905 item 40) (#14519) by @wenshao
• fix(acp-bridge): forget a replaced child's runtime MCP servers (#14527) by @wenshao
• fix(mods): skip catches for re-entered registrations (#14489) by @qqqys
• fix(ci): classify host lock starvation as infrastructure (#14550) (#14552) by @wenshao
• fix(channels/base): strip zero-ink code points from display projections (#14510) by @yiliang114
• fix(managed-agent): preserve rotated credential keys during operator recovery (#14509) by @yiliang114
• fix(managed-agent): verify ambiguous Linux workspace birth times (#14505) by @yiliang114
• fix(core): stamp an aged-out final cron fire with its own minute (#14559) by @yiliang114
• fix(serve): tell a batch its Session can never hold at once to send fewer calls (#14575) by @wenshao
• fix(managed-agent): wait out a wake pump pass before settling its input (#14574) by @wenshao
• fix(core): quiet the crawler's ripgrep probe and stop double-reporting one failed spawn (#14573) by @yiliang114
• fix(managed-agent): bound retention root and head waits (#14558) by @yiliang114
• fix(managed-agent): free Workspace storage after a refused release (#14568) by @wenshao
• test(e2e): widen the channel-worker sandbox budgets (#14587) by @wenshao
• fix(managed-agent): reindent the retired-holder close stage for checkstyle (#14607) by @yiliang114
• test(e2e): mount the fixture root into the container sandbox for workflow ancestor discovery (#14589) (#14593) by @qwen-code-dev-bot
• test(cli): give the Hosted MCP cancellation gates' waitFor an explicit timeout (#14614) by @qwen-code-dev-bot
• fix(managed-agent): accept MariaDB numeric booleans in CHECK validation and repair drain-chain indentation (#14615) (#14618) by @qwen-code-dev-bot

Performance

• perf(managed-agent): avoid redundant session root materialization (#14557) by @yiliang114

Documentation

• Design docs now describe the DingTalk permission field as a checkbox group with single-choice submission and fix a broken reference. (#11774) by @L4XB
• The README Ecosystem section now lists the qwen-audio-agent realtime voice frontend for connecting through ACP stdio mode. (#8630) by @x-lixu
• Adds documentation for looot as a supported MCP web search service, including CLI and settings.json setup. (#13999) by @walidboulanouar
• docs(memory): plan extraction baseline before input changes (#14368) by @yiliang114
• docs(managed-agent): require one child Workspace setting on every replica (#14411) by @wenshao
• docs(managed-agent): stop saying no lifecycle operation sets failure_code (#14422) by @wenshao
• docs(managed-agent): keep hooks on snapshot children (#13905 item 37) (#14440) by @wenshao
• docs(managed-agent): state the read-profile step-3 acceptance (#13905 item 45) (#14450) by @wenshao
• docs: clarify safe mode tool schemas for small contexts (#14445) by @cenab
• docs(managed-agent): specify the M6 engine registration and workspace-change contract (#14479) by @wenshao
• docs(managed-agent): record M6 as implemented in #14483 (#14491) by @wenshao
• docs(managed-agent): add the FG12b and FG12e fault-gate designs (#14447) (#14560) by @wenshao
• docs(managed-agent): format the FG12b design pair (#14580) by @wenshao

Internal Changes

• test(managed-agent): add the FG8e/FG8f monitor_run and background-shell fault gates (#14120) by @wenshao
• Uses the permission probe to skip chmod-based CLI tests only where permission bits are not enforced (#14271) by @wenshao
• Reuses standard utilities for Git output, XML escaping, and artifact byte-range parsing without changing existing behavior (#14307) by @doudouOUC
• Managed Agent server and Runtime Broker sources are consolidated into one module without changing public APIs or behavior. (#14277) by @doudouOUC
• Qwen Code now separates the CLI launcher from the JavaScript interpreter used by hooks, sandbox file writes, and code-mode execution. (#14036) by @chiga0
• Runtime Broker now determines Harness ownership by Runtime Session for shared Workspace bindings, preparing for shared workspace-resident tool runtimes. (#14339) by @doudouOUC
• Internal tests for hosted member commands now stay within fork timeout budgets and avoid unnecessary relay holds. (#14061) by @wenshao
• Adds internal acceptance tests for real extension processes, MCP servers, and Hooks in managed-agent sessions. (#14357) by @doudouOUC
• test(managed-agent): add the FG9a/FG9f session_message and session_command relay fault gates (#14164) (#14333) by @wenshao
• test: skip the chmod-based core, wecom and scripts tests where permission bits do not bind (#14283) by @wenshao
• refactor(managed-agent): Decouple tool publication storage from OSS (#14370) by @doudouOUC
• test(managed-agent): give WorkspaceCreationSchemaMySqlIT its own disposable catalog (#14375) by @wenshao
• test(managed-agent): expect legacy creation to refuse a CSI storage alias (#14382) by @wenshao
• test(managed-agent): add the FG8a/FG8b/FG8d monitor_run and background-shell fault gates (#14163) (#14334) by @wenshao
• test(managed-agent): add the FG9d/FG9e session relay cancellation and SSE-gap fault gates (#14164) (#14402) by @wenshao
• test: gate the remaining chmod-based root skips on the permission-bit probe (#14403) by @wenshao
• test(managed-agent): add the FG8c process-crash gates for monitor_run and background shells (#14163) (#14407) by @wenshao
• test(managed-agent): ground the FG8a/FG8b/FG8d gate choreography in Linux truth (#14398) by @wenshao
• test(managed-agent): keep the harness daemon log in Hosted shell lifecycle evidence (#14401) by @yiliang114
• test(serve): widen the Hosted Harness idle wait on loaded runners (#14394) by @yiliang114
• test(managed-agent): give FG8d Sessions a 10 s writer lease so drain-release gets a capture slot (#14425) by @wenshao
• test(managed-agent): add the FG10a workflow child_run digest-integrity fault gate (#14165) (#14390) by @wenshao
• test(managed-agent): add the FG9b/FG9c session relay store-failure and crash fault gates (#14164) (#14420) by @wenshao
• test(sdk-java): read the heartbeat scheduler queue on its own thread (#14427) by @yiliang114
• refactor(integration-tests): extract the shared helpers duplicated across the hosted gate drivers (#14147) (#14439) by @wenshao
• test(managed-agent): pin the read-profile Session's serve wiring (#13905 items 41-44) (#14459) by @wenshao
• test(managed-agent): add the FG10b/FG10f workflow relay lost-reply and delivery fault gates (#14165) (#14468) by @wenshao
• test(managed-agent): add the FG10c/FG10e workflow relay store-failure and cancellation fault gates (#14165) (#14473) by @wenshao
• ci(sdk-java): run the Hosted failover arms in their own job (#14467) by @wenshao
• test(managed-agent): pin that reconcile_run leaves a live wake turn's run alone (#14480) by @wenshao
• test(managed-agent): add the FG12a MCP reply-loss fault gates (#14447) (#14499) by @wenshao
• test(managed-agent): hold every Hosted IT class to the shared fork budget (#14148) (#14521) by @wenshao
• test(managed-agent): add the FG13c Hook owner-restart fault gates (#14449) (#14490) by @wenshao
• test(managed-agent): add the FG12c MCP restart fault gates (#14447) (#14524) by @wenshao
• test(managed-agent): add the FG13d Hook cancel and timeout fault gates (#14533) by @wenshao
• test(managed-agent): drive the FG12c configure recovery past running (#14447) (#14535) by @wenshao
• test(managed-agent): add the FG11a automation lost-reply fault gates (#14446) (#14534) by @wenshao
• test(managed-agent): pin the reconcile_run route's own crash repair (#14529) by @wenshao
• test(managed-agent): add the FG13d stopped-worker cancel gate (#14541) by @wenshao
• test(managed-agent): add the FG12d MCP release fault gate (#14447) (#14549) by @wenshao
• test(managed-agent): cover workspace-change request validation (#14525) by @wenshao
• test(core): witness the sealed Managed commit-proof clause (#14514) by @yiliang114
• test(managed-agent): add the FG11b Spring-side automation store-fault gates (#14446) (#14551) by @wenshao
• test(managed-agent): keep a workspace-scope permission change out of a worktree session (#14554) by @wenshao
• test(managed-agent): fix the FG12d post-merge evidence and ledger P3s (#14553) by @wenshao
• refactor(web-shell): drop the input-history nav state nothing consumed (#14543) by @yiliang114
• test(managed-agent): add the FG11b Harness-side automation store-fault gates (#14446) (#14564) by @wenshao
• test(managed-agent): drop the FG13d hook-cancel reply through a broker proxy (#14555) by @wenshao
• test(managed-agent): add FG12b Hosted MCP store fault gates (#14569) by @wenshao
• test(managed-agent): add the FG12f MCP stream-gap gates (#14592) by @wenshao
• test(managed-agent): drop the FG13a hook-execute reply through the broker proxy (#14556) by @wenshao
• test(managed-agent): add the FG11c Harness-crash fault gates (#14446) (#14601) by @wenshao
• test(managed-agent): witness the FG12d release ordering (#14447) (#14624) by @wenshao

New Contributors

• @x-lixu made their first contribution in #8630
• @walidboulanouar made their first contribution in #13999
• @cenab made their first contribution in #14445
• @4ekuct25 made their first contribution in #13676

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.25.2...v0.25.3

0.25.2 - 2026-10-15

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• feat(managed-agent): H4e-b2 the lead's mailbox over Session messages (#13892) by @wenshao
• A new private extension imports Vitest 3.2 JSON reports and exposes failures, errors, counts, and diagnostics through MCP tools with search and pagination. (#13854) by @qqqys
• Opt-in headless Mods now support TypeScript authoring, relative imports, typed non-secret configuration, and dependency validation before startup. (#13890) by @qqqys
• Adds a default-off option allowing authorized public REST and WebShell sessions to use foreground Shell with mandatory approval. (#13674) by @doudouOUC
• feat(managed-agent): Introduce the EventTransport seam behind the after-commit publisher (#13909) by @wenshao
• feat(managed-agent): Stop Hosted team members with request_shutdown (H4e-b3a) (#13923) by @wenshao
• Opt-in headless Mod runtime gains ordered middleware dispatch, command interception, and bounded failure recovery without resetting module state. (#13915) by @qqqys
• Adds a Test Failure Explorer app that lets users browse, search, page through, and copy analysis requests for up to 20 selected test failures. (#13916) by @qqqys
• Enables H3 Monitor and background Shell capabilities, allowing background runs to outlive their turn and output artifacts. (#13889) by @wenshao
• Adds a Java mirror of the managed-event-envelope contract with derivation, validation, deduplication, and fixture tests, without changing runtime behavior. (#13940) by @wenshao
• Adds a selection menu to completed Web Shell assistant replies with Copy plain text and Copy Markdown actions while preserving selected formatting. (#13833) by @ytahdn
• Adds Web Shell task reminders for tool approval or AskUserQuestion answers, with notification controls, authorization prompts, and in-panel help. (#13832) by @ytahdn
• Enables opt-in Mods in trusted headless main sessions to intercept tool calls, rewrite arguments, refuse calls, return synthetic results, and check permissions. (#13950) by @qqqys
• feat(managed-agent): verify tool publications asynchronously (#13654) by @doudouOUC
• feat(web-shell): add Managed child Agent task panel (#13936) by @doudouOUC
• feat(managed-agent): recover Hosted G3 parked turns (#13777) by @doudouOUC
• feat(runtime): add private CSI file worker with local-process isolation (#13899) by @doudouOUC
• feat(managed-agent): member board tools ride the session_command channel end to end (H4e-b3b) (#13928) by @wenshao
• feat(web-shell): add read-only Excel artifact previews (#13664) by @ytahdn
• feat(managed-agent): H4 workflow child runtime (#13858) by @wenshao
• feat(managed-agent): member-to-member messages ride the session_command channel (H4e-b3c) (#13955) by @wenshao
• feat(managed-agent): give a Monitor auto-stop its remote exit evidence (#13896) (#13929) by @wenshao
• feat(web-shell): Reuse standard UI for Managed Agents (#13973) by @doudouOUC
• feat(managed-agent): add L4 Shell and MCP session lifecycle (#13931) by @doudouOUC
• feat(core): add a convergence reminder for read-only exploration (#13601) by @yiliang114
• feat(cli): add repeatable --disable-mcp-server to exclude servers per invocation (#14003) by @yiliang114
• feat(web-shell): Add workspace file browser and previews (#13878) by @ytahdn
• feat(managed-agent): add durable team plan approval (H4e-b3d) (#13974) by @wenshao
• feat(managed-agent): Add the LiteTopic journal-event transport (MQ2b) (#14029) by @wenshao
• feat(web-shell): bridge trajectory history and live lifecycle state (#13952) by @qqqys
• feat(web-shell): pulse the header environment entry while tasks run (#14048) by @ytahdn
• feat(web-shell,desktop): update Qwen Code logo to the current brand mark (#14052) by @pomelo-nwu
• feat(managed-agent): Add the public-event transport and the wake/materialization consumers (MQ2c) (#14080) by @wenshao
• feat(managed-agent): consolidate persistence into a 32-table baseline (#14069) by @doudouOUC
• feat(web-shell): open context details in the side panel from the ring (#14086) by @pomelo-nwu
• feat(runtime-broker): add bounded retention for terminal JDBC history (#13642) by @doudouOUC
• feat(web-shell): review MCP App messages before submission (PR3a) (#13958) by @qqqys
• feat(managed-agent): H4f public task cancel for workflow tasks (#14076) by @wenshao
• feat(mods): support selected Claude tool input contracts (PR3b) (#14027) by @qqqys
• feat(sdk-java): add private CSI Session and native SQL authority (#13996) by @doudouOUC
• feat(managed-agent): publish an experimental reference distribution (#14176) by @yiliang114
• feat(web-shell): show a reduced header on the new-task page (#14159) by @ytahdn
• feat(web-shell): upload files by dropping into the file tree (#14145) by @ytahdn
• feat(web-shell): open workflow agents in the detail panel (#14175) by @qqqys

Bug Fixes

• Runtime Broker now rejects closed or deleted managed-agent sessions with a non-retryable error, and closed Hosted Harness connectors fail immediately. (#13330) by @wenshao
• Child worktree runs now settle correctly after merges and stopped runs discard their worktrees, preventing stale workspace state. (#13872) by @wenshao
• Monitor admission and initial-call startup failures are now classified tool errors with full guidance instead of false success. (#13460) by @GoldArowana
• CLI now strips a leading UTF-8 BOM when reading MCP configuration files, including project .mcp.json and --mcp-config file input. (#13596) by @Bug-killer-prog
• Uses the server-reported context limit when retrying after a context-overflow rejection, making reactive compaction recover requests more reliably. (#13788) by @yiliang114
• Fixes a hang when closing Hosted event streams and canceling managed-agent tasks after a Hosted Harness restart, so cancellation completes promptly. (#13877) by @wenshao
• Improves CI routing and safety for the SDK Java MariaDB test lane on trusted triggers. (#13627) by @qwen-code-dev-bot
• Allows idle local managed-agent workspace sessions with original command Hooks to finish retirement after Broker restart, worker exit, or reboot. (#13673) by @doudouOUC
• Makes Claude MCP config import tolerate a leading UTF-8 BOM so configs saved by Windows editors can be imported. (#13711) by @Bug-killer-prog
• fix(managed-agent): e2e runner and image robustness from the #12692 R2 review (#13344) by @wenshao
• fix(managed-agent): return 404 for unknown tool publications (#13602) by @ZedingZhang
• fix(serve): name Hosted recovery-refusal branches in cold-load 409s (#13276) by @wenshao
• fix(cli): bound managed function-hook module evaluation and keep hold-fenced owners recoverable (#13243) by @wenshao
• Fixes a managed-agent publication issue by ensuring placement is selected before retention, quota, and session writer locks. (#13894) by @doudouOUC
• fix(managed-agent): settle provably never-started dispatches at session open (#13801) (#13917) by @wenshao
• Fixes /review Test Plan checks to correctly parse comma-grouped test counts such as 2,239 passed. (#13881) by @wenshao
• Fixes managed-agent config and API behavior with input-size aggregate limits, removed dead config, corrected defaults, and better duration handling. (#13335) by @wenshao
• Localizes Web Shell voice dictation feedback in English and Simplified Chinese, adds dismissible error and no-speech messages, and restores the microphone for retry. (#13938) by @ytahdn
• Repairs CLI CI tests by draining wake park before teardown and restoring shell gate guard expectations. (#13959) by @qwen-code-dev-bot
• fix(lsp): load .lsp.json files saved with a UTF-8 BOM (#13946) by @Shizoqua
• fix(desktop): preserve bundled runtime integrity in AppImage (#13695) by @hahahahahayesyeseys
• fix(release): anchor desktop release notes at the previous desktop release (#13960) by @Jay0130-a
• test(web-shell): re-measure the transcript entry ceiling (#13982) (#13985) by @qwen-code-dev-bot
• fix(cli): report a spawn failure of the CLI child instead of exiting silently (#13077) by @drakeo338
• fix(ci): double the sdk-java MariaDB lane's per-host lock budget (#13989) (#13991) by @qwen-code-dev-bot
• fix(cli): drop the stale gate-close from the restored-shell-gate test (#13983) by @yiliang114
• fix(core): clean the folded cause before getErrorMessage caps the retained subagent error text (#14008) by @yiliang114
• fix(cli): load TOML custom commands saved with a UTF-8 BOM (#13945) by @Shizoqua
• fix(ci): recover reused runner checkouts and Java lock waits (#14017) by @yiliang114
• fix(core): strip a leading BOM before converting a TOML command (#13975) by @Bug-killer-prog
• fix(cli): give bare ↑/↓ back to input history in Virtualized History (#14016) by @chiga0
• fix(web-shell): hide the discontinued built-in OAuth model in Settings (#13971) by @Bug-killer-prog
• fix(cli): window the OpenTUI transcript to fix the blank-screen resume (#13669) by @chiga0
• fix(cli): keep the OpenTUI composer caret steady (#13693) by @chiga0
• fix(managed-agent): share title blankness and pin the lease-renewal clamp (#13979) by @yiliang114
• fix(managed-agent): count http/function result Hooks at child admission (#13978) by @yiliang114
• fix(cli): stop OpenTUI slash commands reading their own dispatch as busy (#12960) by @MohammedAlkindi
• fix(cli): recognize markdown fences only on matching lines (#13331) by @GoldArowana
• fix(sdk): strip terminal escapes from plaintext tool previews (#12751) by @sxh313
• fix(cli): split comma-separated --include-tools/--exclude-tools in mcp add (#12985) by @yanyukang2003-ux
• fix(web-shell): Ignore standalone whitespace assistant chunks (#14031) by @ytahdn
• fix(managed-agent): order a sender's commands by commit sequence, and make the member restart IT interrupt a live wait (#14014) by @wenshao
• fix(desktop): let Web Shell receive OS file drops (#14053) by @pomelo-nwu
• fix(managed-agent): numeric parity and family-list hoisting for publication families, with end-to-end admission and wedge witnesses (#13815) by @wenshao
• fix(cli): reject batch plan targets that alias the same path (#14055) by @yiliang114
• fix(managed-agent): close shell profiles after native file calls (#14045) by @doudouOUC
• fix(managed-agent): recover lost lifecycle detach replies (#14039) by @doudouOUC
• fix(serve): resume committed file results after runtime retirement (#14059) by @doudouOUC
• fix(release): raise the npm package unpacked-size budget to 160 MiB (#14066) by @qwen-code-dev-bot
• test(managed-agent): widen process-death observation windows in the operator recovery probe (#14072) (#14079) by @qwen-code-dev-bot
• fix(managed-agent): passively reattach child relay sessions after control-plane restart (#14020) by @yiliang114
• fix(desktop): keep desktop-latest release notes pointing at the current release (#13900) by @shpaker
• test(managed-agent): handle the procfs exit race in the operator recovery probe (#14090) by @wenshao
• test(serve): widen hosted approval inline-limit windows for CI contention (#14084) (#14091) by @qwen-code-dev-bot
• fix(ci): raise the Hosted MySQL fork and step budgets past the L4 class (#14125) (#14128) by @qwen-code-dev-bot
• test(managed-agent): Pin the history response size boundary (#14111) by @loglapa
• fix(cli): bound short OpenTUI scroll dialogs (MYWO-3, MYWO-7) (#14034) by @chiga0
• fix(managed-agent): Repair the Hosted storage-guard fixtures red since #14069 (#14131) by @yiliang114
• fix(acp): preserve cancellation intent across session recovery (#13436) by @yiliang114
• test(managed-agent): register storage for Hosted admission gates (#14154) by @yiliang114
• test(managed-agent): stabilize retraction snapshot verification (#14151) by @yiliang114
• fix(managed-agent): settle journal-only parked Harness turns on takeover (#13847) (#14078) by @wenshao
• test(managed-agent): wait for the published Snapshot before naming its Items (#14156) by @yiliang114
• fix(managed-agent): Wire OSS capture budgets into Hosted sessions (#14124) by @doudouOUC
• test(managed-agent): cover coordinator-only Hosted restart (#14155) by @yiliang114
• fix(build): re-apply the ink patch when install skipped lifecycle scripts (#14170) by @yiliang114
• fix(cli): preserve Unicode queries in at completion (#13866) by @dakjdakd
• fix(sdk-python): resolve Windows command shims on the effective PATH (#13862) by @dakjdakd
• fix(cli): gate non-interactive /update installs on enableAutoUpdate (#13665) by @fengyue-xve
• fix(core): preserve fuzzy edit line boundaries (#13484) by @GoldArowana
• fix(sdk): release reconnect transports created after disposal (#13864) by @dakjdakd
• fix(sdk-java): get the MariaDB lane past #13996's schema, recovery IT and baseline witness (#14177) by @wenshao
• fix(core): resolve ${CLAUDE_PLUGIN_ROOT} in plugin markdown without hooks (#14194) by @yiliang114
• fix(ci): repair Java backlog and schema/Hosted test failures (#14178) by @yiliang114
• fix(ci): re-pin Java gate CI tests to the daily-trigger shape (#14223) (#14227) by @qwen-code-dev-bot
• fix(sdk-java): restore session-root materialization in acquireWriter (#14224) (#14228) by @qwen-code-dev-bot
• fix(core): keep Artifact guidance out of record_artifact description (#14206) by @zjgzx1988
• fix(managed-agent): settle the monitor loop when its resume finds a settled record (#14074) by @wenshao
• fix(ci): repair daily failure reporting and contract guards (#14238) by @yiliang114
• fix(ci): repin the sdk-java test job's MAVEN_ARGS count for the checkstyle gate (#14247) (#14258) by @qwen-code-dev-bot
• fix(sdk-java): Order the MQ3 gate's imports for the new checkstyle rules (#14253) by @wenshao
• fix(core): select the DashScope provider for Model Studio workspace domains (#14245) by @qqqys
• fix(managed-agent): close the critical R2 review findings on #12692 (#13325) by @wenshao
• fix(core): surface the provider request id on bodyless API errors (#14133) by @yiliang114
• fix(managed-agent): hand one run's session messages over in send order (#14235) by @wenshao
• fix(Managed agent): settle never-started background dispatches in the live turn (#14021) by @yiliang114
• fix(core): keep the git snapshot out of the system prompt (#14255) by @qqqys
• fix(ci): gate Desktop and Java SDK releases on checks of the release commit (#14259) by @yiliang114
• fix(serve): align Hosted team delivery answers (#14267) by @yiliang114
• fix(ci): drain Hosted test turns and repair nightly browser checks (#14275) by @yiliang114
• fix(cli): close post-merge takeover findings from the #13083 review (#13188) by @wenshao
• fix(vscode): move conversation payloads out of globalState (#14285) by @yiliang114

Documentation

• Adds design documentation for managed-agent team member commands, plan approval, and shutdown on the Hosted path. (#13912) by @wenshao
• docs(managed-agent): Resolve the EventTransport §5.1 LiteTopic verification ledger (#13927) by @wenshao
• docs(managed-agent): Resolve the EventTransport §11 routing and wake-carrier questions (#13933) by @wenshao
• Adds documentation planning the EventTransport MQ2 Java producer/consumer wiring into four independently reviewable sub-slices. (#13939) by @wenshao
• docs(sdk): clarify empty core-tool allowlists across SDKs (#13949) by @Zhangchenhao-111
• docs: record local tool visibility comparison and verification steps (#13767) by @yiliang114
• docs(readme): simplify evaluation overview and refine trend chart (#14093) by @pomelo-nwu
• docs(cli): stop naming Terminal.app in the arrowKeysScrollTranscript description (#14092) by @yiliang114
• docs(runtime-broker): design Workspace-resident Managed Tool Runtime (#14214) by @doudouOUC
• docs: remind to update workflow meta-tests when changing CI workflows (#14239) by @yiliang114
• docs(design): drop stray conflict markers from the actor-roles design (#14270) by @doudouOUC

Internal Changes

• Adds internal FG7 fault-gate tests that verify managed-agent channel behavior during control-plane and provider failures. (#13839) by @wenshao
• Improves test stability for child-Workspace Git layout validation by avoiding a possible directory-name collision. (#13844) by @qwen-code-dev-bot
• test(integration): sync the updatedInput ask approval on the dialog, not the tool row (#13849) (#13855) by @qwen-code-dev-bot
• Raises the Web Shell transcript entry size ceiling after selection-copy strings increased bundle size, with no production behavior change. (#13964) by @qwen-code-dev-bot
• test(runtime-broker): budget the 1 MiB tool round trips past the request deadline (#13990) by @yiliang114
• test(managed-agent): clean up recovery fixtures after probe failures (#14018) by @yiliang114
• test(managed-agent): stabilize publication and API contract fixtures (#14012) by @yiliang114
• test(managed-agent): land the deferred findings from the definition-execution review (#13980) by @yiliang114
• test(core): cover unfinished invocation guard before turn reclamation (#13257) by @jayfeng7
• refactor(core): one record/array gate for file-history snapshot payloads (#13981) by @yiliang114
• test(managed-agent): deflake the transient verification retry rescan (#14037) (#14042) by @qwen-code-dev-bot
• Add SerpApi to web search MCP services list (#13104) by @pulkitchowdry
• test(managed-agent): fix the shell lifecycle recovery journal-head assertions (#14103) (#14108) by @qwen-code-dev-bot
• test(serve): widen hosted recovery waitFor windows for CI contention (#14101) (#14107) by @qwen-code-dev-bot
• test(managed-agent): Run the EventTransport fault rows against a real RocketMQ broker (MQ2d) (#14109) by @wenshao
• test(integration): Cover a completing write_file tool call over ACP (#14144) by @yiliang114
• test(managed-agent): retry the team-lead hosted boot past the free-port race (#14153) by @yiliang114
• refactor(core): give the Code Mode tool-result stamp a single owner (#13666) by @fengyue-xve
• test(ci): pin the main-failure watcher's whole job condition (#14233) by @wenshao
• chore: adopt qwen-code-checks.xml and fix all checkstyle violations (#14186) by @doudouOUC
• test(managed-agent): Run the EventTransport fault matrix against two nodes (MQ3) (#14242) by @wenshao
• test(cli): skip the chmod-based tests where permission bits do not bind (#14162) by @wenshao
• refactor(runtime-broker): name the local-process kind and Workspace isolation policy (#14254) by @doudouOUC
• test(managed-agent): gate G3 Step 3 recovery, model-round takeover and control-plane fencing (#14181) by @wenshao
• diagnostics(ci): capture RocketMQ latency stages separately (#14241) by @yiliang114
• refactor(managed-agent): Reuse Apache Commons utilities (#14280) by @doudouOUC

New Contributors

• @ZedingZhang made their first contribution in #13602
• @Zhangchenhao-111 made their first contribution in #13949
• @hahahahahayesyeseys made their first contribution in #13695
• @Jay0130-a made their first contribution in #13960
• @drakeo338 made their first contribution in #13077
• @MohammedAlkindi made their first contribution in #12960
• @jayfeng7 made their first contribution in #13257
• @yanyukang2003-ux made their first contribution in #12985
• @shpaker made their first contribution in #13900
• @pulkitchowdry made their first contribution in #13104
• @loglapa made their first contribution in #14111
• @dakjdakd made their first contribution in #13866
• @fengyue-xve made their first contribution in #13665

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.25.1...v0.25.2

0.25.1 - 2026-10-11

Highlights

See the complete change list below.

Breaking Changes

No known breaking changes.

Complete Change List

Features

• Restores the selected text-only connection and validated session or workspace route after Android Activity recreation, while preserving explicit retry screens after failures. (#12247) by @jabrailkhalil
• Improves mobile profile screen accessibility with clearer connection actions, editor labels, validation announcements, and scrollable storage recovery. (#12129) by @jabrailkhalil
• Adds an experimental Kubernetes CSI-backed runtime that validates workspaces, persists pod identity, and records durable worker acknowledgements for managed sessions. (#13289) by @doudouOUC
• Adds background Shell and Monitor capabilities to the Managed Agent runtime, enabling durable background shell sessions and task event reporting. (#13265) by @wenshao
• Web Shell /btw side can now create independent side-task conversations in a parent session's secondary workspace, including while the parent is responding. (#13468) by @wenshao
• Agent and Goal coordination tools are now deferred by default and discovered on demand, reducing unnecessary tool declarations in ordinary tool-call sessions. (#13033) by @yiliang114
• Local Managed Agent runtime tool outcomes are now durably recorded, including parameters, receipts, and checkpoints, before results are returned to the model. (#13291) by @wenshao
• Hosted approval requests can now include bounded previews of file and shell input, exposed through public and WebShell APIs with truncation metadata. (#13400) by @yiliang114
• Lets creators change a Workspace-bound Managed Session's working directory through a durable, idempotent operation. (#13247) by @wenshao
• Adds read-only glob file discovery to Hosted Workspace /2 profiles while keeping searches inside the session working directory. (#13166) by @yiliang114
• Adds planned Stage H5 agent channel API route shapes and persistence tables for channel routes and outbound deliveries. (#13497) by @wenshao
• feat(web-shell): take back a cancelled prompt that produced nothing (#13488) by @wenshao
• feat(hosted): show captured inputs in native approval cards (#13407) by @yiliang114
• feat(managed-agent): EventTransport message-envelope contract (#13498) by @wenshao
• feat(managed-agent): adopt the next Hosted Harness generation (G3) (#13174) by @wenshao
• feat(managed-agent): enumerate the public and WebShell surface with an admission gate (#13543) by @wenshao
• feat(managed-agent): H4a child agent and child acceptance record contract (#13505) by @wenshao
• feat(managed-agent): give Hosted turns the Workspace's project context (#13168) by @yiliang114
• feat(managed-agent): H6a schedule and automation run record contract (#13536) by @wenshao
• feat(managed-agent): H5a channel route and delivery record contract (#13548) by @wenshao
• feat(web-shell): Add fullscreen viewing for history tables (#13589) by @wenshao
• feat(managed-agent): prove Shell process-group stops with a worker ledger (M5c) (#13352) by @wenshao
• feat(managed-agent): add W1c offline workspace migration (#13260) by @doudouOUC
• feat(agents): session-centric multi-agent collaboration (#13467) by @yiliang114
• feat(managed-agent): add reliable ACTIVE Workspace deletion (L3) (#13354) by @doudouOUC
• feat(hooks): apply PreToolUse updatedInput with full revalidation (#13442) by @qqqys
• feat(web-shell): add trajectory search and diagnostic filters (#13264) by @qqqys
• feat(web-shell): restore aggregated workspace hosts in sidebar (#13677) by @yiliang114
• feat(managed-agent): persist workspace actor roles and session owners (#13544) by @wenshao
• feat(managed-agent): H4b child Session runtime (#13550) by @wenshao
• feat(web-shell): browse earlier trajectory windows (#13716) by @qqqys
• feat(agents): remove the thread backend and run A2A on sessions (#13583) by @yiliang114
• feat(trajectory): preserve model request execution identity (#13740) by @qqqys
• feat(managed-agent): H5b/H5c channel runtime for the email reference adapter (#13572) by @wenshao
• feat(managed-agent): add opt-in production replay-floor advancement for Session event retention (#12380) (#13621) by @qwen-code-dev-bot
• feat(managed-agent): H4c workflow child kind and child launch budget (#13754) by @wenshao
• feat(managed-agent): enforce workspace actor roles across the bound-Session surface (#13545) by @wenshao
• feat(managed-agent): H6b/H6c automation runtime for persistent definitions (#13598) by @wenshao
• feat(core): record prompt execution context (#13712) by @ytahdn
• feat(managed-agent): H4d-a session message record contract and child continuation rules (#13786) by @wenshao
• feat(managed-agent): H4e-a team record contract (#13811) by @wenshao
• feat(managed-agent): child Workspace capability for managed child Sessions (#13753 I1) (#13781) by @wenshao
• feat(trajectory): record and replay request lifecycle (PR6B) (#13755) by @qqqys
• feat(managed-agent): restart-recoverable foreground child wait (fixes #13708) (#13769) by @wenshao
• feat(web-shell): connect saved daemons simultaneously, no-flash host switching (#13727) (#13739) by @yiliang114
• feat(managed-agent): H4e-b1 lead-side team runtime (#13824) by @wenshao
• feat(managed-agent): H4f public task cancel for child agent tasks (#13823) by @wenshao
• feat(trajectory): record native tool execution lifecycles (#13836) by @qqqys
• feat(managed-agent): worktree admission for managed child agents (#13753 I2) (#13841) by @wenshao
• feat(managed-agent): H4d-b session message runtime (#13822) by @wenshao
• feat(mods): run experimental headless Mod commands (#13843) by @qqqys
• feat(managed-agent): enable the H4e-b1 team domains after the physical pass (#13846) by @wenshao
• feat(managed-agent): Collect retired stream-capture tool outputs (#13534) (#13554) by @wenshao
• feat(managed-agent): execute pinned AgentDefinition revisions (#13530) by @yiliang114
• feat(core): Keep complete file reads on one opened source (#13776) by @doudouOUC
• feat(managed-agent): support WebShell session cwd changes (#13760) by @doudouOUC
• feat(web-shell): add Send now for queued mid-turn messages (#13681) by @wenshao

Bug Fixes

• Replaces a selected remote host in Web Shell without losing agent bindings by revoking the old credential and moving bindings to the new host. (#13430) by @yiliang114
• Adds a 60-second wait before retrying E2E build artifact downloads to reduce false failures from immediate back-to-back attempts. (#13292) by @qwen-code-dev-bot
• Makes yamllint and shellcheck CI lanes fail clearly when the git file list is empty instead of silently passing or showing misleading errors. (#12650) by @qwen-code-dev-bot
• Improves managed-agent SSE delivery by using Condition-based waiting instead of pinning virtual-thread carriers. (#13402) by @wenshao
• Prevents managed-agent harness attachment creation from blocking concurrent map monitors, avoiding virtual-thread carrier pinning under cold bursts. (#13403) by @wenshao
• Fixes /context detail so per-tool MCP rows are scaled consistently and add up to the MCP tools row on both estimate and provider-count paths. (#13390) by @Shizoqua
• Aligns memory-index prompt loading with the writer budget so oversized entries are omitted whole and later complete entries remain available in order. (#13315) by @beibeihk
• Fixes vim-mode clipboard paste on Windows by making clipboard reads more reliable with PowerShell UTF-8 output, a longer timeout, and normalized line endings. (#13198) by @Bug-killer-prog
• Stops automatic QWEN.md and AGENTS.md discovery at the Git root so files in parent directories above the repository no longer enter requests. (#13381) by @agammann
• Extension Git operations no longer stall waiting for interactive credentials; unauthenticated sources now fail immediately. (#13454) by @yiliang114
• Settings file replacement now preserves the existing file's POSIX permission bits under stricter umask, improving safety and consistency when saving settings. (#13293) by @doudouOUC
• Fixes WeChat login QR code generation by sending required iLink headers with the request. (#13482) by @yiliang114
• Recognizes llama.cpp context-overflow wording so compaction runs instead of looping on HTTP 400. (#13421) by @yiliang114
• Recovers complete function-style XML tool calls so they execute instead of remaining assistant text. (#13437) by @yiliang114
• Honors memory.agentMaxTurns in user-scoped memory dream and restricts turn/timeout limits to trusted operator settings. (#13462) by @yiliang114
• Prevents MCP permission rules for one server from authorizing a different colliding server. (#12531) by @yiliang114
• Fixes web-shell managed session UI correctness issues including stale error banners, turn settling, pending-prompt handling, session switching, and provider rebuild behavior. (#13342) by @wenshao
• Makes main-branch CI failures caused only by GitHub infrastructure issues rerun failed jobs once instead of filing an autofix issue. (#13512) by @qwen-code-dev-bot
• fix(core): keep nested tool_call arguments open on Responses; name the target in bridged validation errors (#12901) by @yiliang114
• Ensures memory agent budget settings are honored only from User, System, and SystemDefaults scopes, with workspace values ignored and warned. (#13508) by @yiliang114
• Prevents an XML tool call quoted inside a parameter value from being dispatched as a real tool call. (#13515) by @yiliang114
• fix(ci): raise the Hosted verify step ceiling above its failsafe fork timeout (#13503) (#13504) by @qwen-code-dev-bot
• fix(ci): sweep stale pool residue before sdk-java checkouts (#13507) by @qwen-code-dev-bot
• fix(managed-agent): check the replay before publishing a domain record (#13376) by @wenshao
• fix(core): key the models.dev catalog under dotted as well as dashed ids (#13299) by @yiliang114
• fix(mobile): resolve document providers once per selection (#13147) by @jabrailkhalil
• fix(managed-agent): close the three Critical H0c follow-ups from #13300 (#13355) by @wenshao
• fix(core): stop advertising unsupported LSP dynamic registration (#13494) by @shenyankm
• fix(core): stop JSONL prefix reads when their budget is met (#13486) by @GoldArowana
• fix(web-shell): escape control characters in the managed approval args preview (#13549) by @yiliang114
• fix(core): budget side-query output tokens against the resolved context window (#13244) by @yiliang114
• fix(agent): enforce per-model concurrency cap on foreground sub-agents (#12461) by @Andrea-Bruno
• fix(runtime-broker): keep the HTTP exchange cancel off the deadline completion path (#13562) by @wenshao
• fix(ci): skip millisecond latency budgets on the release ECS pool (#13541) (#13547) by @qwen-code-dev-bot
• fix(docs): apply Prettier formatting to the daemon event schema (#13573) (#13580) by @qwen-code-dev-bot
• fix(acp): preserve branch checkpoints after Code Mode turns (#13574) by @yiliang114
• fix(web-shell): allow deleting queued prompts during admission (#13594) by @yiliang114
• fix(serve): invalidate the Hosted Workspace context on instruction-file changes (#13605) by @yiliang114
• fix(web-shell): follow an accepted approval answer to its terminal operation (#13609) by @yiliang114
• fix(managed-agent): commit oversized message records as inline chunks (#13375) by @wenshao
• fix(ci): classify runner-acquisition failure as infrastructure (#13622) (#13626) by @qwen-code-dev-bot
• fix(web-shell): sanitise model-supplied text at the approval card's sibling render sites (#13578) by @yiliang114
• fix(feishu): preserve text and clean up failed inbound file writes (#13337) by @jabrailkhalil
• fix(managed-agent): clear successful Workspace migration diagnostics (#13655) by @doudouOUC
• fix(managed-agent): stop a bound Turn under refused authorization (#13163) by @yiliang114
• fix(acp): include stored attachment paths in model context (#13668) by @ytahdn
• fix(core): preserve GLM vision metadata for dashed model IDs (#13651) by @yiliang114
• fix(goals): separate aggregate summaries from external facts (#13575) by @yiliang114
• fix(core): hide the hook child console on Windows (#13670) by @yiliang114
• fix(acp-bridge): keep local published file artifacts ephemeral (#12395) by @zjgzx1988
• fix(workflow): show results and failures in /workflows run details (#13646) by @qqqys
• fix(core): gate discovery hints on registered capabilities (#13576) by @yiliang114
• fix(memory): preserve the prompt prefix when memory indexes change (#13521) by @yiliang114
• fix(core): surface PreToolUse ask content on MCP tool confirmations (#13687) (#13697) by @yiliang114
• fix(web-shell): preserve artifact cards during reconnect (#13714) by @ytahdn
• fix(web-shell): allow downloading changed workspace artifacts (#13718) by @ytahdn
• fix(browser-use): fail fast when no Native Messaging host can be registered (#13699) by @yiliang114
• fix(web-shell): Select worktree mode without extra confirmation (#13607) by @wenshao
• fix(web-shell): show workspace artifacts by their filename (#13672) by @zjgzx1988
• fix(cli): preserve complete Hosted no-tool model Parts (#13750) by @doudouOUC
• fix(ci): raise the O4 gate step ceiling above its failsafe fork timeout (#13684) (#13688) by @qwen-code-dev-bot
• fix(verify-pr): drop the stale "ships no JDK" claim from the Java lane guidance (#13791) (#13792) by @qwen-code-dev-bot
• fix(release): reclaim docker disk and gate the data root before the sandbox image build (#13479) (#13481) by @qwen-code-dev-bot
• fix(managed-agent): unblock H3 background Shell and Monitor start path (#13789) by @wenshao
• fix(core): leave unknown ${identifier} placeholders literal in subagent prompts (#13694) by @yiliang114
• fix(mcp): discover tools on the readResource lazy-spawn path (#13806) by @yiliang114
• fix(cli): match ink's OpenTUI popup geometry and completion truncation (#12559) by @chiga0
• fix(ci): bound E2E sandbox image cleanup (#13686) by @zonemeen
• fix(ci): reject unsupported Bash in workflow size checks (#13470) by @zonemeen
• fix(web-shell): collapse background results with turn steps (#13825) by @ytahdn
• fix(managed-agent): answer journal-only calls on every interrupted-turn route (#13829) by @wenshao
• fix(core): suppress trailing orphan thinking tags in prose (#13600) by @yiliang114
• fix(core): make lazy-spawn MCP discovery best-effort (#13828) by @yiliang114
• fix(browser-use): report Windows as unsupported instead of advising an install (#13675) by @yiliang114
• fix(managed-agent): observe background-process exits at Session re-acquire (#13830) by @wenshao
• fix(permissions): escalate repeated destructive-command denials to manual approval (#13636) by @yiliang114
• fix(core): surface failed or unavailable LSP diagnostics as errors (#13128) by @yiliang114
• fix(managed-agent): harden the managed panel failure lifecycle and worker path containment (#13179) by @wenshao
• fix(core): close Managed session correctness gaps from #12693 post-merge review (#13332) by @wenshao
• fix(session): reserve retained snapshot prompt identities on resume (#13729) by @yiliang114
• fix(runtime): retain CSI mount identity through directory resolution (#13772) by @doudouOUC
• fix(managed-agent): reconcile approval delivery and concurrent session titles (#13682) by @yiliang114
• fix(hosted): Count future result-hook mounts at child admission (#13773) by @yiliang114
• fix(managed-agent): settle the deferred review findings of PR #13352 (#13630) by @wenshao
• fix(core): tell the parent model why a foreground subagent stopped (#13624) by @yiliang114

Performance

• Speeds up full extension scans by loading extension directories in bounded parallel batches while preserving order and error behavior. (#13379) by @callmeYe

Documentation

• Adds Stage H4/H5/H6 slice design documents for managed agents, channels, and automation. (#13499) by @wenshao
• Adds English and Chinese design documentation for the optional cross-node EventTransport based on message queues. (#13500) by @wenshao
• Repairs managed-agent documentation findings from the R2 review, fixing stale links, port guidance, stage labels, and incorrect claims. (#13343) by @wenshao
• docs(serve): fix stale permission-route note and document pending-prompt events (#13281) by @LUOSENGWA
• docs(managed-agent): add the delivery ledger for #12380 (#13565) by @wenshao
• docs(troubleshooting): add a "connection resets on some networks" entry (TLS-stack-selective middleboxes) (#13235) by @LUOSENGWA
• docs(managed-agent): bring the delivery ledger up to main d0ddd020c8 (#13593) by @wenshao
• docs(serve): fix Prettier formatting in the daemon event schema (#13585) (#13587) by @qwen-code-dev-bot
• docs(managed-agent): link every PR and issue reference in the delivery ledger (#13631) by @wenshao
• docs(core): point the session-agents contract at its SDK mirror (#13648) by @yiliang114
• docs(verify-pr): add local-round publish, follow-up, mutation and rig rules (#13756) by @wenshao
• docs(managed-agent): record H3 Linux acceptance blockers and replay inputs (#13764) by @yiliang114
• docs(verify-pr): name scenario axes in scope selection and attribute merge-only reds (#13813) by @wenshao
• docs(managed-agent): H4e-b1 lead-side team runtime design (#13817) by @wenshao
• docs(managed-agent): Name RocketMQ LiteTopic (5.5.0+) as the preferred P2 EventTransport candidate (#13826) by @wenshao
• docs(verify-pr): key the turn-axes rule to the diff, not the central claim (#13819) by @wenshao
• docs(managed-agent): H4e-b2 team mailbox design (#13879) by @wenshao
• docs(verify-pr): budget the turn-axes secondary A/B as a harness slot (#13850) by @wenshao

Internal Changes

• Adds post-merge test coverage and hygiene checks for HTTP store integrity scenarios, including multi-page journals and restore refusal on inconsistent data. (#13341) by @wenshao
• Stabilizes the /hooks dialog E2E test by using CI-aware timeouts and asserting against the rendered dialog. (#13427) by @qwen-code-dev-bot
• Increases the waitFor timeout for hosted tool-turn approval tests to reduce flaky CI failures. (#13399) by @qwen-code-dev-bot
• test(integration): make the utf-bom defaultFileEncoding test deterministic (#13196) by @Bug-killer-prog
• test(managed-agent): pin untested contracts from the #12692 R2 review (#13348) by @wenshao
• test(integration): let the headless workflow-skill E2E inherit the suite timeout (#13518) (#13520) by @qwen-code-dev-bot
• test(managed-agent): commit authority-valid deltas in the restore byte-budget test (#13551) by @wenshao
• test(core): pin the limit-less alias shape the models.dev projection ships (#13539) by @yiliang114
• test(integration): require a successful monitor tool call (#13452) by @GoldArowana
• test(managed-agent): harden pinning witnesses and add renewal-arm pinning witness (#13388 follow-up) (#13401) by @wenshao
• test(integration): share the Hosted proxy header filter across all store relays (#13431) by @wenshao
• test: isolate hosted catalog refresh and verify ACP defaults (#13653) by @yiliang114
• test(desktop): pin both Linux release legs to the ubuntu-22.04 glibc floor (#13616) by @tonydzi
• test(managed-agent): tolerate scheduling slack in the detach heartbeat window (#13658) by @yiliang114
• test(integration): let workflow-completion reply waits inherit the helper default (#13552) (#13555) by @qwen-code-dev-bot
• refactor(transcript): centralize internal Code Mode result detection (#13577) by @yiliang114
• test(memory): cover request-tail catalog and protocol contracts (#13736) by @yiliang114
• test(managed-agent): verify Hosted Shell operator recovery on MySQL and MariaDB (#13759) by @yiliang114
• test(managed-agent): cover AgentDefinition review follow-ups (#13757) by @yiliang114
• ci(sdk-java): re-check open PRs for Flyway version collisions as main moves (#13770) by @wenshao
• ci(verify): provision a Java toolchain for sdk-java PRs (#13763) by @doudouOUC
• refactor(managed-agent): classify child Session run kinds by name (#13779) by @wenshao
• test(managed-agent): widen HostedHarnessMySqlIT lifecycle budgets for CI MySQL stalls (#13780) (#13783) by @qwen-code-dev-bot
• test(cli): bind ephemeral loopback port in container-mode tests (#13775) (#13778) by @qwen-code-dev-bot
• ci(sdk-java): stop the API contract version from regressing or stalling (#13804) (#13808) by @wenshao
• test(cli): avoid port conflicts in managed runtime container tests (#13569) by @zonemeen
• test(cli): pin the background capture pause across the launcher-exit flush (#13834) by @wenshao
• test(managed-agent): run the H4e-b1 team flow end to end in the Hosted MySQL lane (#13867) by @wenshao

New Contributors

• @beibeihk made their first contribution in #13315
• @agammann made their first contribution in #13381
• @GoldArowana made their first contribution in #13486
• @LUOSENGWA made their first contribution in #13281
• @zonemeen made their first contribution in #13686

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.25.0...v0.25.1

0.21.2 - 2026-07-31

Highlights

• Autofix now defers lower-severity suggestions after five rounds and posts visible notices when refusing to proceed due to round limits. (#7913, #8067)
• Added an opt-in Auto Recall profile that automatically searches external context with built-in security limits. (#7877)
• Enhanced the /verify command with increased time budgets, evidence screenshots, and new maintainer verification techniques. (#8010, #8014, #8016)
• Web Shell now supports contextual task panels, theme-aware visuals, and correctly handles artifact previews and session states. (#7929, #8098, #8078, #8106)
• GitHub channels now emit exactly one final comment per event and display eyes reactions to indicate active agent processing. (#8033, #8061)
• Added APIs and UI controls to view, approve, and revoke channel pairing approvals within specific workspaces. (#8045, #8081)

Breaking Changes

No known breaking changes.

Complete Change List

Features

• Autofix now defers lower-severity suggestions after five change rounds, allowing only critical findings and failed checks to drive further code modifications. (#7913) by @qqqys
• Goal v3 is now connected to the turn engine with deterministic permit handling, ensuring runtime continuations do not consume session or recursive turn budgets. (#7895) by @qqqys
• Static image reads now return a canonical, auto-oriented JPEG overview with source dimensions to support reliable zooming within shared resource budgets. (#7911) by @qqqys
• The triage system now mechanically triggers the 2b-bis sandboxed-lane recommendation when a Stage 2 draft explicitly contains a not-verified sentence. (#7965) by @wenshao
• Adds an opt-in Auto Recall profile that automatically searches external context when you submit a prompt, with built-in security limits. (#7877) by @doudouOUC
• Updates WebShell to use a shared renderer for streaming charts, ensuring canonical markdown-chart blocks work without extra host configuration. (#7916) by @zhangxy-zju
• Improves ripgrep reliability by retrying failed searches and correctly distinguishing between no matches found and incomplete search results. (#7888) by @water-in-stone
• feat(channels): add DingTalk interactive cards (#6930) by @BenGuanRan
• feat(web-shell): refine advanced table controls (#7999) by @ytahdn
• Updated triage comments to lead with a clear pass or fail verdict and display English by default with Chinese translations folded. (#7974) by @wenshao
• feat(triage): run external /verify on ECS behind a workspace wipe (#7985) by @wenshao
• Added the security.allowPrivateNetworkHooks setting to allow HTTP hooks to access private network addresses in trusted platform-managed environments. (#7968) by @xurik
• Added a customizable action slot to split pane headers that automatically overflows into a menu when space is limited. (#7808) by @samuelhsin
• Added a GitLab polling channel adapter that converts GitLab todos into inbound messages for automated processing. (#7862) by @OrbitZore
• Added APIs and SDK helpers to list approved pairing senders and revoke access within a specific workspace. (#8045) by @qqqys
• Introduced an optional reasonFilter setting for GitHub channels to skip unwanted notification reasons and prevent unintended agent actions. (#8031) by @xianjianlf2
• Review approvals with zero findings on non-trivial diffs now explicitly disclose low signal status in the verdict line. (#7987) by @wenshao
• Added a threshold setting to automatically preload deferred tools into the context window when their total size fits within limits. (#7922) by @DragonnZhang
• feat(review): add review run — headless review with a machine-readable verdict (#7983) by @wenshao
• Enhanced the verify-pr skill with seven new techniques from maintainer verification rounds to improve change validation. (#8010) by @wenshao
• Increased the /verify agent budget from 25 to 120 minutes to allow comprehensive verification matching local maintainer rounds. (#8014) by @wenshao
• Added automatic milestone summaries every tenth takeover round to provide visibility into long-running fix attempts. (#8046) by @wenshao
• Adds a temporary eyes reaction to GitHub issues and comments while an agent turn is running to indicate active processing. (#8061) by @yiliang114
• Reduces scan delays by skipping inspection of idle autofix candidates to prioritize fresh engagements. (#8049) by @wenshao
• The /verify command now successfully captures and includes evidence screenshots in its reports. (#8016) by @wenshao
• Ensures maintainer feedback containing specific phrases is not deferred in Critical-only mode by implementing a per-source feedback budget. (#8071) by @wenshao
• Tags UserPromptSubmit hook context in model requests and records display provenance to ensure session replays accurately reflect what the model saw. (#7956) by @zjgzx1988
• Adds byte-cursor paging to large text file reads across HTTP, ACP, and SDK surfaces to efficiently retrieve subsequent content without rescanning. (#8002) by @doudouOUC
• The qwen review test-efficacy command now detects unprotected safety statements using statement-level mutation probes that delete single lines to verify test coverage. (#8020) by @wenshao
• feat(web-shell): add contextual task panels (#7929) by @ytahdn
• Subagents using the fork type now support an optional fork_tools allowlist to restrict execution to specific canonical tool names or MCP server patterns. (#8066) by @destire-mio
• The Channel editor now allows operators to view, approve, and revoke pairing approvals with explicit confirmation dialogs distinct from the configured allowlist. (#8081) by @qqqys
• Enhanced Web Shell visuals with theme-aware composer highlights, interactive dot animations for empty sessions, and a new typewriter effect for placeholders. (#8098) by @ytahdn

Bug Fixes

• Interactive end-to-end tests now scale read-then-write waits with the environment timeout instead of using a fixed 15-second limit to prevent flakiness. (#7943) by @qwen-code-dev-bot
• CI now cleans stale .qwen/ directories before checkout to prevent permission denied errors caused by restrictive file modes left by previous jobs. (#7977) by @qwen-code-dev-bot
• CI verify jobs now correctly restore write permissions on the .qwen tree before returning workspace ownership to prevent checkout failures on self-hosted runners. (#7992) by @yiliang114
• Fixes an issue where ECS runners on specific hosts failed to update the Qwen binary to the requested version. (#8000) by @yiliang114
• Fixes directory listing to clearly distinguish between folders that are empty and those that were skipped due to item budget limits. (#7868) by @chinesepowered
• Fixes a bug where code blocks in Feishu messages could exceed size limits by incorrectly calculating space needed for closing fences. (#7850) by @chinesepowered
• Ensures standalone WebShell URLs preserve authentication tokens and deployment base paths when navigating between sessions. (#7926) by @ytahdn
• fix(web-shell): polish Channel pairing requests (#7997) by @qqqys
• fix(cli): correct hardware cursor off-by-one in fullscreen mode (#7998) by @chiga0
• fix(cli): prevent SGR mouse events from being swallowed as paste on Windows (#7988) by @chiga0
• fix(serve): allow bounded reads of large text files (#7947) by @doudouOUC
• fix(web-shell): prefer artifact type metadata in cards (#7973) by @ytahdn
• fix(web-shell): reduce composer input latency (#8015) by @ytahdn
• Fixed session history pagination to prevent failures when loading earlier records by using stable record boundaries instead of opaque cursors. (#8001) by @ytahdn
• Fixed release note generation to correctly handle cases where the previous release tag diverges from the current release target. (#7970) by @qwen-code-dev-bot
• Increased the browser daemon SDK bundle size budget to 177KB to accommodate recent feature additions and restore successful builds. (#8024) by @wenshao
• Fixed keyboard handling in Kitty protocol terminals to correctly map the Command or Super key modifier to the meta flag for shortcuts. (#7996) by @chiga0
• Improved diagnostic messages to distinguish between a missing channel worker and a worker that is intentionally draining during workspace reloads. (#7932) by @destire-mio
• Added preflight disk space checks to the build-test command to prevent installation and build failures caused by insufficient storage. (#7986) by @wenshao
• Fixed Ctrl+Shift+C to correctly copy text in the terminal instead of triggering quit or clear actions. (#8011) by @chiga0
• Artifact file writes are now automatically recorded upon success, eliminating the need for a separate manual registration step. (#7914) by @chiga0
• Fixed an issue where pressing Enter during MCP prompt completion incorrectly treated optional parameters as mandatory. (#7995) by @qwen-code-dev-bot
• Disabled native cron jobs in daemon-managed Channel sessions to ensure scheduled tasks are correctly tracked and delivered. (#8034) by @qqqys
• GitHub channels now fail startup with guidance if configured with a self-only operator allowlist that cannot receive usable notifications. (#8055) by @zjunothing
• Removed a schema combinator from the send_message tool to fix compatibility with Anthropic models that previously rejected the request. (#7989) by @netbrah
• Fixed a terminal resize issue that caused excessive scrolling and history re-rendering during panel toggle animations. (#8009) by @chiga0
• Forked background agents now correctly refresh tool capabilities and system instructions when resumed while preserving conversation history. (#7927) by @DragonnZhang
• GitHub channel publications now emit exactly one final comment per event, preventing duplicate or intermediate outputs while recording detailed audit metadata for every attempt. (#8033) by @yiliang114
• The review-pr job now restores workspace ownership before checkout to prevent permission denied errors on reused self-hosted runners. (#8062) by @yiliang114
• Preview version generation now detects existing stable releases and bumps the patch number to prevent npm publishing conflicts for channel packages. (#7978) by @yiliang114
• Daemon-managed session writers now use an integrity-protected handoff protocol with sealed locks to ensure safe ownership transfer during managed shutdowns. (#7976) by @doudouOUC
• Transport stream retries are now allowed during the thinking-only phase by distinguishing non-thought content chunks from intermediate processing steps. (#7938) by @ComplexSimply
• The reasonFilter setting now validates input arrays, rejects unknown reasons, and treats empty arrays as unset to ensure reliable notification routing. (#8035) by @yiliang114
• Fixed an issue where CJK-heavy text could cause output token limits to be exceeded, preventing API errors during generation. (#7963) by @zambalee
• Ensured active Todo lists persist correctly across tool turns to maintain context without interfering with user input priority. (#7919) by @yiliang114
• Fixed environment variable propagation to ensure skill subprocesses use the correct CLI build and model configuration. (#7993) by @wenshao
• Fixed the autofix loop to properly count timeouts toward failure limits and corrected status messages during handoffs. (#8044) by @wenshao
• Fixes takeover acknowledgments by posting confirmations directly from the command to prevent event loss and long delays. (#8043) by @wenshao
• Stabilizes web shell table controls by fixing column sizing, frozen-column shadows, and scroll position during row expansion. (#8041) by @ytahdn
• Makes the interactive read-then-write integration test deterministic by using a fake server instead of a live LLM. (#8064) by @qwen-code-dev-bot
• Adds a visible Queued on server status in the web shell for messages waiting in the daemon queue before execution. (#8065) by @wenshao
• Fixes web shell artifact previews by enabling scripts in sandboxed iframes and correctly handling binary image files. (#8078) by @ytahdn
• Optimized the Comment Attachment Guard to skip unnecessary runner allocation, reducing queue congestion for hosted jobs. (#8095) by @wenshao
• Web Shell now correctly translates legacy /skills invocations with arguments into direct skill commands. (#8103) by @ytahdn
• Updated the verify-pr agent to account for evidence capture costs within its execution budget to ensure images are generated. (#8104) by @wenshao
• Web Shell now displays a clear failure status and Retry action for prompts that fail before daemon admission. (#8106) by @ytahdn
• Added retry capability and in-progress states to Web Shell question submissions to prevent silent failures. (#8096) by @ytahdn
• Fixed Web Shell to correctly distinguish and label one-time tool approvals that also switch the session to Default mode. (#8099) by @ytahdn
• Autofix now posts visible notices to PRs when refusing to proceed due to round-cap limits instead of logging only. (#8067) by @wenshao
• Prevents loss of verified autofix results by merging concurrent pushes and retrying when the PR head moves during a long agent run. (#8042) by @wenshao
• Properly aborts active dynamic workflow runs during session shutdown to prevent orphaned dispatches and unnecessary token usage. (#8107) by @qqqys
• Stabilizes the setModel E2E test by detecting turn completion via result messages instead of counting assistant messages. (#8075) by @qwen-code-dev-bot
• Increases the per-turn model response timeout in multi-model E2E tests from 30s to 60s to reduce flakiness on CI. (#8111) by @qwen-code-dev-bot
• Prevents repeated workspace skill rescans by making status reads side-effect free and implementing generation-guarded caching with ETag support. (#8080) by @doudouOUC
• Daemon session maintenance now uses a writer-lease protocol to isolate transcript updates and prevent conflicts during session deletion, archiving, and cleanup operations. (#7975) by @doudouOUC
• Fixed a build failure by routing side-task rollback operations through the workspace session service to resolve an unbound identifier introduced by recent maintenance changes. (#8144) by @wenshao
• Web Shell worktree sessions now consistently execute commands in the session's effective working directory and defer MCP discovery until trusted relocation completes. (#8068) by @wenshao
• Permission-control E2E tests now use shared response timeouts to prevent false failures caused by slow model responses during dynamic permission mode changes. (#8149) by @wenshao
• Fixed a flaky end-to-end test for auto-edit permissions to ensure deterministic results without relying on live model behavior. (#8154) by @wenshao
• Fixed session details menus overflowing on narrow layouts and added a keyboard-accessible action to copy the full session ID. (#8127) by @dreamWB

Internal Changes

• The asyncGenerator canUseTool test now validates tool-call arguments instead of strict file content equality to stabilize assertions against model wording variations. (#7939) by @qwen-code-dev-bot
• Adds an isolated CI pipeline to automatically run SWE-bench benchmarks on stable Qwen Code releases. (#7656) by @DennisYu07
• Migrated flaky end-to-end tests to a deterministic fake server to eliminate failures caused by model output variance and latency. (#7934) by @yiliang114
• Expanded regression tests now verify session history pagination behavior, including active record tracking and edge cases for page-size parsing. (#7907) by @PratikWayase
• Opt-in daemon benchmarks now measure immediate prompt dispatch stages, including HTTP acceptance, message echo, and queue-wait durations. (#7994) by @doudouOUC
• Stabilizes a flaky E2E test case for subagent delegation by refining prompt instructions without changing production code. (#8073) by @qwen-code-dev-bot

New Contributors

• @xianjianlf2 made their first contribution in #8031
• @zambalee made their first contribution in #7963

Full Changelog: https://github.com/QwenLM/qwen-code/compare/v0.21.1...v0.21.2

0.16.1 - 2026-05-23

Added

• telemetry: Phase 4a — TTFT capture + GenAI semconv dual-emit (#3731) (#4417)

Fixed

• core,cli: close tool_use↔tool_result invariant across all failure paths (#4176)
• vscode: skip redundant tsc build in prepackage to prevent TS5055 (#4401)
• core: preserve tab-indented notebook formatting (#4373)
• scripts: renormalize CRLF storage for install-qwen-standalone.bat (#4427)
• build: tree-shake React reconciler dev build to prevent PerformanceMeasure leak (#4462)
• cli: stabilize flaky sticky-todo remeasure test (#4416)
• cli: gate mintty OSC 8 detection on TERM_PROGRAM_VERSION ≥ 3.3 (#4420) (#4451)
• release: move constants above entry point to avoid TDZ error (#4398)

Other

• chore(deps): update express from 4.21.2 to 5.2.1 (#4458)

0.16.0 - 2026-05-21

Added

• cli: wrap markdown links in OSC 8 so wrapped URLs stay clickable (#4037)
• cli: support batch deletion of sessions in /delete (#3733)
• subagents: use fastModel for Explore subagent (#4086)
• perf: progressive MCP availability — MCP no longer blocks first input (#3994)
• core: strip inline media before chat compaction summary (#4101)
• tools: add generic worktree support — EnterWorktree/ExitWorktree + Agent isolation (#4073)
• cli: add ModelScope as a built-in third-party API provider (#4150)
• core: add image+video support for Qwen3.6-35B-A3B quant variants (#4106)
• hooks: Add TodoCreated and TodoCompleted hooks for todo lifecycle events (#3378)
• hooks: add prompt hook type with LLM evaluation support (#3388)
• core,cli: add generic atomicWriteFile, wire into Write/Edit tools, upgrade @types/node (#4096)
• cli: warn users that rewind is disabled in IDE mode (#4122)
• cli: argument hint + --auto completion for /rename (#4048)
• cli: add baseline /doctor memory diagnostics (#4180)
• cli: add session-scoped /goal command with judge-driven turn continuation (#4123)
• rewind: add file restoration support to /rewind command (#4064)
• skills: add /stuck diagnostic skill for frozen sessions (#4133)
• telemetry: unify span creation paths for hierarchical trace tree (#4126)
• cli: readline Ctrl+P/N for history and selection navigation (#4082)
• cli: add built-in status line presets with interactive dialog (#4120)
• cli: add fork-session resume flag (#4159)
• telemetry: add interaction span and detailed sensitive attributes (#4097)
• core: PR-2.5 — post-promote stream redirect + natural-exit registry settle (#3831 follow-up) (#4102)
• cli: add configurable plansDirectory for Plan Mode (#4062)
• cli: add structured memory diagnostics JSON (#3785)
• core: fail impossible goals (#4230)
• serve: add /demo debug page for qwen serve daemon (#4132)
• worktree: Phase C — session persistence, hooksPath, Footer + WorktreeExitDialog, three-mode --resume restore (#4174)
• core: extend cross-auth fast models to agents (#4153)
• cli,core: add Auto approval mode with LLM classifier (#4151)
• cli: per-turn /diff with interactive dialog (#4277)
• cli: add session path status command (#4124)
• core: inject git status into system prompt and refine Explore/git-log guidance (#4110)
• core: add NotebookEdit tool for Jupyter notebooks (#3900)
• cli: respect /editor preference in Ctrl+X external editor (#4310)
• telemetry: Phase 2 — tool.blocked_on_user + hook spans (#3731) (#4321)
• installer: add standalone hosted install and uninstall flow (#3828)
• telemetry: support custom resource attributes and add metric cardinality controls (#4367)
• skills: support priority field in SKILL.md for sorting skill display order (#4155)

Changed

• cli: revert dynamic slash command LLM translation (#4145)
• core: TaskBase envelope + foreground subagent persistence (#3970)
• auth: unify provider config in core, simplify /auth as "Connect a Provider" (#4287)
• core: undo x-api-key + Authorization double-emit (#4342) — regresses IdeaLab-style proxies (#4385)

Fixed

• core: normalize cumulative OpenAI stream deltas to suffixes (#3896)
• cli: auto-restore prompt and preserve queue on cancel (#4023)
• core: tag subagent OpenAI JSON logs (#4099)
• dashscope: use URL hostname check instead of regex to avoid ReDoS (CodeQL) (#4112)
• core: improve runtime fetch options error handling and documentation (#3997)
• telemetry: address PR #3847 review follow-ups for trace correlation (#4058)
• search: make empty-query exit synchronous and normalize Windows Backspace (#3981)
• anthropic: allow cache_control on tool_result blocks (#4121)
• core: merge IDE context into user prompt (#3980)
• cli: apply /language output to running session without restart (#4143)
• core: correct context-usage Footer for prompt size and Anthropic caches (#4109)
• core: support cross-auth fast side queries (#4117)
• vscode: preserve thinking state and recover missing edit snapshots (#4147)
• cli: handle MinTTY Ctrl+Backspace as delete-previous-word (#4059)
• cli: preserve debug session across sandbox relaunch (#4060)
• hooks: inject SessionStart additionalContext into chat context (#4115)
• i18n: Correct zh-TW translations to match Traditional Chinese conventions (#4129)
• core: refresh systemInstruction in setTools() so progressive MCP tools reach the model (#4166)
• vscode-ide-companion: use existing editor group for diff instead of forcing a new one (#4130)
• core: add heap-pressure auto-compaction safety net (#4186)
• cli: pass rewind selector test props (#4211)
• lsp: expose status and startup diagnostics (#3649)
• rewind: restore upstream TOCTOU ordering + heal sticky failed marker (#4216)
• test: clear boundedPromise timers to prevent unhandled rejections in abort-and-lifecycle test (#4220)
• ui: trim background task results and show newest first (#4094) (#4125)
• core: align shell tool description with configured shell (#4170)
• cli: include skill base dir in slash commands (#4224)
• cli: restore ACP prompt counter on resume (#4233)
• core: extend DashScope provider detection with additional hostname rules (#4157)
• core: apply tool name migrations at dispatch (#4213)
• cli: record mid-turn queued user prompts (#4215)
• add cache limits to prevent OOM during build/test (#4188)
• core: preserve read-before-write state across idle microcompaction (#4243)
• telemetry: Phase 1.5 polish — fallback order, abort-as-result, log/span consistency (#4302)
• cli: /status preserves prior error history items (#4169) (#4265)
• core: decouple auto-memory recall from main-agent request path (#4172)
• core: apply defaultModalities() on env-var-only model config (#4219) (#4262)
• cli: block Windows Tab approval-mode toggle when input has a Tab consumer (#4308)
• core: mirror Qwen3 reasoning on outbound history (#4294)
• test: count result messages instead of assistant messages in multi-model E2E test (#4341)
• test: raise timeout for Windows installer end-to-end tests (#4352)
• review: harden SKILL.md against weak-model rule skipping (#4340)
• cli: remove QWEN_OAUTH gate from feedback dialog (#4316)
• core: replace structuredClone with shallow copy to prevent OOM in long sessions (#4286)
• core: align session hook matcher targets (#4354)
• core: handle MiMo tool-result media (#4281)
• core: deduplicate geminiChat recovery continuation text (#3966)
• ci: resolve TS5055 release build failure since May 19 (#4383)

Performance

• cli: code-split lowlight to cut startup V8 parse cost (#4070)

Documentation

• auth: add custom API key wizard PRD (#3583)
• user + design docs for --json-schema structured output (#4051)

Other

• ci(deps): bump docker/* actions to Node 24 majors (silences GitHub Node 20 deprecation warning) (#4131)
• test(integration): pin simple-mcp-server to legacy MCP path until #4163 is fixed (#4164)
• chore(deps): re-upgrade ink 6 → 7.0.3 (upstream Static remount fix landed) (#4119)
• Add stop hook blocking cap (#4208)
• [codex] Allow custom output directory for /export (#4193)
• test(perf): skip daemon baseline harness under sandbox (#4234)
• test: reduce wait-dependent UI test delays (#3987)
• chore(vscode): run development ACP CLI from source (#4283)
• Support active goal stream events and non-interactive goals (#4273)
• Pin fetch to bundled undici for undici higher versions compatibility (#4238)
• chore: add .github/release.yml to support skip-changelog label (#4327)
• Expose active goal in stream JSON (#4314)

0.15.11 - 2026-05-13

Added

• cli: core built-in i18n coverage (#3871)
• core: write runtime.json sidecar for active sessions (#3714)
• telemetry: inject traceId/spanId into debug log files for OTel correlation (#3847)
• tools: defer low-frequency built-in tools to reduce initial prompt size (#4022)
• installer: add standalone archive installation (#3776)
• cli: Ctrl+B promote keybind (#3831 PR-3 of 3) (#3969)
• cli: add --json-schema for structured output in headless mode (#3598)
• skills: Add codegraph skill for PR review risk analysis and conflict detection (#3910)
• tools: keep ask_user_question always-visible to surface clarification UX (#4041)
• core: improve Anthropic proxy compatibility and enable global prompt cache scope (#4020)
• cli: add tools.toolSearch.enabled setting for prefix-caching models (#4069)
• core: replace fdir crawler with git ls-files + ripgrep fallback (#3214)
• dashscope: support DASHSCOPE_PROXY_BASE_URL for prompt cache via API gateway (#3991)
• telemetry: add hierarchical session tracing spans (#4071)

Changed

• cli: remove legacy qwen auth CLI subcommand, redirect to /auth TUI dialog (#3959)
• core: route side-query LLM calls through runSideQuery chokepoint (#3775)
• telemetry: remove dead useCollector setting and unreachable TelemetryTarget.QWEN (#4061)
• deps: downgrade ink 7 → 6 to fix Static-remount TUI regression from #3860 (#4083)

Fixed

• cli: keep long model stats header on one line (#4032)
• test: repair stale --json-schema integration assertion (#4075)
• cli: improve rendering on narrow terminals (#3968)
• channels: expand tilde in channel cwd config (#4045)
• cli: preserve table ANSI color across wrapped lines (#4050)
• core: log internal OpenAI JSON requests (#4081)

Performance

• core: bound session-list metadata reads to head/tail 64KB; pool buffer; lazy message count (#3897)

Documentation

• telemetry: align config and docs semantics for target, outfile, and CLI flags (#4066)

Other

• test: stabilize main e2e flakes (#3992)
• ci: skip unnecessary release and SDK checks (#3984)
• chore(deps): upgrade ink 6.2.3 → 7.0.2 + bump Node engine to 22 (#3860)
• chore(core): runtime.json sidecar follow-ups from #3714 review (#4030)
• Upgrade GitHub Actions for Node 24 compatibility (#1876)
• doc[sdk-python] Expand Python SDK usage documentation (#3995)
• ci(e2e): stabilize MCP/CLI flows and cancel stale main runs (#4039)

0.15.10 - 2026-05-10

Added

• core: add reactive compression on context overflow (#3879)
• memory: add autoSkill background project skill extraction (#3673)
• cli: improve slash command discovery (#3736)
• core: support QWEN_HOME env var to customize config directory (#2953)
• vscode: add message edit/rewind and message metadata UI (#3762)
• add /diff command and git diff statistics utility (#3491)
• tools: add ToolSearch for on-demand loading of deferred tool schemas (#3589)

Fixed

• cli: validate /model command arguments (#3963)
• core: log the OpenAI request actually sent on the wire (#3767)
• core: drop disabled MCP server from health status registry (#3916)
• core: filter Mistral reasoning content at request boundary (#3882)
• cli: preserve comments and formatting in settings.json during migration write-back (#3861)
• cli: unfreeze Ctrl+O compact-mode toggle on long conversations (#3905)
• cli: replace clearTerminal with targeted repaint on resize (#3967)
• core: harden reactive compression follow-ups (#3985)
• core: throttle shell tool live text updates (#3902)
• core: unify Edit/WriteFile prior-read with Claude Code; close #3964 + #3945 (#4002)

Other

• test(cli): drop wait-dependent SessionPicker search tests (closes #3977) (#3978)
• [codex] fix monitor notifications for subagents (#3933)
• feat(telemetry) suppress OpenTelemetry diagnostics from UI (#3986)

0.15.9 - 2026-05-08

Added

• telemetry: add sensitive span attribute opt-in (#3893)
• add commit attribution with per-file AI contribution tracking (#3115)
• sdk-python: replace verbatim release notes inheritance with --generate-notes (#3835)
• cli: add Idealab as third-party provider (#3955)
• session: add /branch to fork the current conversation (#3539)
• core: foreground → background promote integration (#3831 PR-2 of 3) (#3894)
• cli: searchable /resume picker with focus-aware modes (#3880)
• skills: reload slash commands when SkillManager fires change event (#3923)

Changed

• cli: provider-first auth registry with unified install pipeline (#3864)

Fixed

• core: per-agent ContentGenerator view via AsyncLocalStorage (#3707)
• core: accept partial reads in prior-read enforcement (#3932)
• cli,core: live-phase panel-ownership filter + post-delete statusChange emit (#3919)
• core: close bound-tool gap on runForkedAgent's YOLO wrapper (#3892)
• vscode: mark Qwen OAuth coder-model as Discontinued in model picker (#3948)
• cli: show tool details in subagent approval banner (#3956)
• cli: trim blank streaming tails from live preview (#3965)
• core: route countSessionMessages through parseLineTolerant (#3692)

Other

• ci(release): keep skip-ci out of release PR titles (#3950)
• chore: Add bilingual requirement to create-issue command (#3952)
• [codex] Persist ACP model selection (#3947)
• ci: reduce PR test matrix runtime (#3962)

0.15.8 - 2026-05-07

Added

• web-templates: add light theme and toggle to /export HTML (#3908)
• cli: replace inline AgentExecutionDisplay with always-on LiveAgentPanel (#3909)

Fixed

• skills: allow symlinks pointing outside the skills directory (#3915)
• core: foreground agent entry lingering in status bar after completion (#3921)
• cli: prevent ESC in background tasks dialog from cancelling running request (#3922)
• memory: address code review feedback for auto-memory recall (#3866)
• cli: use tmux-safe dots spinner to reduce redraw pressure (#3903)

Other

• test(sdk): align tool-control E2E with prior-read enforcement (#3898)
• ci(issue-followup-bot): render bot comment newlines correctly (#3918)
• ci(release): skip CI on the version-bump squash commit on main (#3912)

0.15.7 - 2026-05-07

Added

• core: add FileReadCache and short-circuit unchanged Reads (#3717)
• core: add shared permission flow for tool execution unification (#3723)
• review: expand review pipeline + qwen review CLI subcommands (#3754)
• telemetry: define HTTP OTLP endpoint behavior and signal routing (#3779)
• core: event monitor tool with throttled stdout streaming (Phase C) (#3684)
• cli: add MCP health pill to footer (#3741)
• cli: wire Monitor entries into combined Background tasks dialog (#3791)
• cli: include monitors in /tasks + add interactive-mode hint (#3801)
• sdk-python: add PyPI release workflow (#3685)
• core: support reasoning effort 'max' tier (DeepSeek extension) (#3800)
• core: hint to background long-running foreground bash commands (#3809)
• skills: parallelize loading + add path-conditional activation (#3604)
• sdk-python: add network timeouts to release version helper (#3833)
• cli: improve export format completion navigation (#3701)
• cli: Add ability to switch models non-interactively from the cli (#3783)
• weixin: add image sending support via CDN upload (#3781)
• core,cli: surface and cancel auto-memory dream tasks (#3836)
• cli: route foreground subagents through pill+dialog while running (#3768)
• core: enforce prior read before Edit / WriteFile mutates a file (#3774)
• cli: customize banner area (logo, title, hide) (#3710)
• core: add signal.reason convention for ShellExecutionService (#3831 PR-1 of 3) (#3842)
• cli: expand TUI markdown rendering (#3680)

Changed

• extract shared release helper utilities (#3834)

Fixed

• cli: honor proxy setting (#3753)
• cli: restore SubAgent shortcut focus (#3771)
• vscode-companion: align package eslint config with root and style cleanup (#3782)
• test: restore abort-and-lifecycle stdin-close test to pre-#3723 version (#3777)
• core: inject thinking blocks for DeepSeek anthropic-compatible provider (#3788)
• cli: stop double-wrapping and double-printing API errors in non-interactive mode (#3749)
• telemetry: suppress async resource attribute warning on startup (#3807)
• core: address post-merge monitor tool and UI routing issues (#3792)
• core: clear FileReadCache on every history rewrite path (#3810)
• core: unescape shell-escaped file paths in Edit, WriteFile, and ReadFile tools (#3820)
• openai: parse MiniMax thinking tags (#3677)
• telemetry: add bounded shutdown timeout and fix service.version resource attribute (#3813)
• acp: run auto compression before model sends (#3698)
• core: coalesce MCP server rediscovery (#3818)
• core: activate skills from discovered result paths (#3852)
• core: use per-model settings for fast model side queries (#3815)
• core: prevent auto-memory recall from blocking main request (#3814)
• sdk-python: standardize TAG_PREFIX to include v suffix (#3832)
• cli: prevent file paths from being treated as slash commands (#3743)
• core: auto-compact subagent context to prevent overflow (#3735)
• core: shrink file diff session records (#3872)
• core: rebuild tool registry on subagent Config overrides so bound tools resolve to the subagent (#3873)
• core: create temp dir before saving truncated shell output (#3875)
• core: improve stream rate-limit retry handling (#3790)
• core: address @tanzhenxin's PR-1 review notes (post-merge follow-up to #3842) (#3886)
• core: stop per-subagent ToolRegistry on foreground-fork path (#3887)
• cli: warn on ignored provider generation config (#3883)

Documentation

• core: point background-shell + monitor guidance at both /tasks and the dialog (#3808)
• cli: document new banner customization settings (#3885)

Other

• chore: remove legacy Gemini workflows (#3725)
• Add background agent resume and continuation (#3739)
• Feat/stats model cost estimation rebase (#3780)
• ci: add Qwen Code issue follow-up bot workflow (#3854)

0.15.6 - 2026-04-30

Fixed

• memory: use project transcript path for dream (#3722)
• cli: bound SubAgent display by visual height to prevent flicker (#3721)
• cli: keep sticky todo panel compact (#3647)
• core: replay DeepSeek reasoning_content on all assistant turns (#3747)
• cli: correct model precedence — argv > settings > auth env vars (#3645)
• core: preserve reasoning_content in rewind, compression, and merge paths (#3579) (#3737)
• cli: persist directory add entries (#3752)
• lsp: 修复 LSP 文档、isPathSafe 限制，并提升 LSP 工具调用率 (#3615)
• vscode-companion: fill slash commands into input on Enter instead of auto-submitting (#3618)
• ci: add merge-back PR for stable releases in release workflow (#3764)

Other

• chore(core): drop tool token usage tracking (#3727)

0.15.5 - 2026-04-29

Added

• core: wire background shells into the task_stop tool (#3687)
• skills: add tmux-real-user-testing skill for readable TUI test logs (#3577)
• cli: wire background shells into combined Background tasks dialog (#3720)

Fixed

• cli: refresh static header on model switch (#3667)
• core: inject reasoning_content on DeepSeek tool-call replays (#3729)

Other

• mcp config as cli (#1279)

0.15.4 - 2026-04-28

Added

• Adds Catalan language support (#3643)
• cli: add API preconnect to reduce first-call latency (#3318)
• cli: Add argument-hint support for slash commands (#3593)
• cli,core: LLM-generated summary labels for tool-call batches (#3538)
• cli: add OSC notification support for iTerm2, Kitty, and Ghostty (#3562)
• vscode: add tab dot indicator and notification system (#3106) (#3661)
• core: model-facing agent control (task_stop, send_message, per-agent transcript) (#3471)
• cli: background-agent UI — pill, combined dialog, detail view (#3488)
• core: managed background shell pool with /tasks command (#3642)

Changed

• config: dedupe QWEN_CODE_API_TIMEOUT_MS env override logic (#3653)

Fixed

• vscode-companion: slash command completion not triggering after message submit (#3609)
• cli: guard gradient rendering without colors (#3640)
• config: support QWEN_CODE_API_TIMEOUT_MS across OAuth and non-OAuth paths (#3629)
• cli: add API Key option to qwen auth interactive menu (#3624)
• core: recover from }{ glued records on session JSONL load (#3606) (#3656)
• core: split tool-result media into follow-up user message for strict OpenAI compat (#3617)
• core: handle shell line continuations in command splitting (#3600)
• cli: recognize OpenAI-compatible providers in qwen auth status (#3623)
• core,cli: stop stripping reasoning on model switch/history load (#3682)
• ci: use squash merge for SDK release auto-merge (#3690)
• cli: preserve description in subject-bearing thought chunks (#3691)
• core: treat ask_user_question multiSelect as optional (#3699)
• core: set DeepSeek V4 context to 1M and output to 384K (#3693)
• ci: preserve preview version overrides (#3705)

Other

• chore(gitignore): add .codex directory (#3665)
• Feat/openrouter auth (#3576)
• test(cli): remove 8 flaky TUI input tests surfaced by CI history mining (#3694)

0.15.3 - 2026-04-26

Added

• vscode: add native context menu copy actions for webview chat (#3477)
• cli: add Traditional Chinese (zh-TW) as a UI language option (#3569)
• vscode: expose /skills as slash command with secondary picker (#2548)
• cli: add conversation rewind feature with double-ESC and /rewind command (#3441)
• adds a Space-to-preview affordance to the /resume session picker (#3605)
• cli: add sticky todo panel to app layouts (#3507)

Changed

• cli: undo OPENAI_MODEL precedence change in modelProviders lookup (#3567) (#3633)

Fixed

• cli: memoize useHistory() return to avoid unnecessary re-renders (#3547)
• cli: respect OPENAI_MODEL precedence in CLI model resolution (#3567)
• cli: add TUI flicker foundation fixes (#3591)
• cli: drain runExitCleanup before process.exit in error handlers (#3602)
• review: respect /language output setting for local reviews (#3611)
• test: update rewind E2E Test 1 assertion after isRealUserTurn fix (#3622)
• core: preserve settings-sourced apiKey when registry model envKey is absent (#3495)
• telemetry: use safeJsonStringify in FileExporter to avoid circular reference crash (#3630)
• core: match DeepSeek provider by model name for sglang/vllm (#3613) (#3620)

Performance

• core: cut runtime sync I/O on tool hot path by 91% (#3581)

Documentation

• github: tighten PR template validation guidance (#3522)
• telemetry: clarify Alibaba Cloud console entry (#3498)

Other

• feat(SDK) Add Python SDK implementation for #3010 (#3494)
• test(arena): cover select dialog key actions (#3614)

0.15.2 - 2026-04-24

Added

• session: auto-title sessions via fast model, add /rename --auto (#3540)
• web-search: remove built-in web_search tool, replace with MCP-based approach (#3502)
• docs: add qwen-code skills, agents, and updated AGENTS.md (#3575)
• vscode-companion: support /export session command (#2592)

Changed

• core: make OpenAI converter stateless (follow-up to #3525) (#3550)
• vscode-ide-companion: undo #3450 split-stream timestamp sharing (#3573)

Fixed

• core: treat empty 'pages' parameter as unset in ReadFile (#3559)
• i18n: sync mismatched keys between en.js and zh.js (#3534)
• cli: remove residual blank lines after MCP init completes (#3509)
• sdk-java: pass custom env to CLI process (#3543)
• cli: promote resubmitted history prompt to most recent (#3531)
• Strengthen error handling in qwenOAuth2.ts to prevent unhandled 'error' event (#3481)
• acp: support SSE and HTTP MCP servers in ACP mode (#3574)
• cli: run ACP Agent tool calls concurrently (#2516) (#3463)
• cli: disable Kitty keyboard protocol on SIGINT to prevent garbled 9;5u output (#3544)
• cli: dispatch queued slash commands through the slash path (#3523)
• core: preserve reasoning_content during session resume and active sessions (GH#3579) (#3590)

0.15.1 - 2026-04-23

Added

• cli: combine elapsed + timeout in shell time indicator (#3512)

Fixed

• core: scope StreamingToolCallParser per stream, not per Converter (#3516) (#3525)
• cli: stop slash completion render loop (#3533)

Other

• chore: bump version to 0.15.1 (#3541)

0.15.0 - 2026-04-22

Added

• acp: add complete hooks support for ACP integration (#3248)
• optimize compact mode UX — shortcuts, settings sync, and safety (#3100)
• hooks: Add HTTP Hook, Function Hook and Async Hook support (#2827)
• memory: managed auto-memory and auto-dream system (#3087)
• cli: support multi-line status line output (#3311)
• skills: add /batch skill for parallel batch operations (#3079)
• background subagents with headless and SDK support (#3076)
• core: add path-based context rule injection from .qwen/rules/ (#3339)
• cli: add dual-output sidecar mode for TUI (#3352)
• bind M-d to a reasonable (Emacs-like) default (#3358)
• core: detect tool validation retry loops and inject stop directive (#3178)
• mcp: add OSC 52 copy hotkey for OAuth authorization URL (#3393)
• vscode-ide-companion: add dedicated agent execution display (#2590)
• cli: add early input capture to prevent keystroke loss during startup (#3319)
• cli: support refreshInterval in statusLine for periodic refresh (#3383)
• core: add dynamic swarm worker tool (#3433)
• tools: add Markdown for Agents support to WebFetch tool (#2734)
• core: enhanced loop detection with stagnation + validation-retry checks (#3236)
• cli: add /doctor diagnostic command (#3404)
• vscode-companion: enable Plan Mode toggle and approval UI (#2551)
• cli: add session recap with /recap and auto-show on return (#3434)
• cli: add bare startup mode (#3448)
• vscode-ide-companion: support /insight command (#2593)
• cli: add slashCommands.disabled setting to gate slash commands (#3445)
• core: PDF text extraction fallback and Jupyter notebook parsing (#3160)
• cli: add OAuth configuration flags to mcp add (#3442)
• cli: add tool execution progress messages (#3155)
• cli: make ACP message rewrite timeout configurable (#3475)
• cli: attribute /stats rows to the originating subagent (#3229)
• webui: render markdown in generic and web-fetch tool outputs (#3469)
• cli: display real-time token consumption during streaming (#2742) (#3329)
• retry: add persistent retry mode for unattended CI/CD environments (#3080)
• vscode: replace OAuth with Coding Plan / API Key provider setup (#3398)
• arena: add comparison summary for agent results (#3394)
• session: add rename, delete, and auto-title generation for session (#3093)
• cli: cap inline shell output with configurable line limit (#3508)
• cli: auto-detect terminal theme ('auto' or unset) (#3460)
• cli: Phase 2 — slash command multi-mode expansion, ACP fixes, and UX improvements (#3377)

Changed

• core: move fork subagent params from execute() to construction time (#3255)
• cli: replace slash command whitelist with capability-based filtering (Phase 1) (#3283)

Fixed

• sdk: avoid leaking process exit listeners in ProcessTransport (#3295)
• cli: prevent statusline spawn EBADF from crashing CLI (#3264) (#3310)
• cli: remember "Start new chat session" until summary changes (#3308)
• cli: defer update notifications until model response completes (#3321)
• core: limit skill watcher depth to prevent FD exhaustion (#3320)
• core: strip thinking blocks from history on model switch (#3315)
• core: add shell argument quoting guidance to prevent special char errors (#3327)
• cli: reduce terminal redraw cursor movement (#3381)
• dingtalk: only suffix '(cont.)' on continuation chunks, not the first (#2977)
• dingtalk: preserve empty text after @mention strip instead of falling back (#2978)
• dingtalk: remove reactionContext map to stop leak on blocked messages (#2979)
• sandbox: fall back to 'latest' tag when image name has no colon (#2962)
• scripts: remove duplicate bundle rmSync in clean script (#2964)
• integration-tests: honor stdinDoesNotEnd option (#2966)
• scripts: Fix "undefined Options: ..." in generated JSON schema for enum settings without descriptions. (#2963)
• text-buffer: unify offset-to-position logic (#2969)
• weixin: check full 4-byte PNG magic signature (#2970)
• cli: re-arm disconnected listener on rebuilt AcpBridge after crash (#2975)
• sdk: settle pending next() promise in Stream.return() to prevent hangs (#2981)
• cli: auto-submit on number key press in AskUserQuestionDialog (#3407)
• tool-registry: add lazy factory registration with inflight concurrency dedup (#3297)
• cli: wait for dual output stream shutdown (#3416)
• build: invoke tsx directly via node --import instead of npx (#3237)
• core: support older Git during repository initialization (#3436)
• cli: /clear dismisses active /btw side-question dialog (#3431)
• cli: let /btw use live conversation context (#3429)
• display ">100%" when context usage exceeds limit (#2766)
• ui: constrain shell output width to prevent box overflow (#2857)
• core: remove abort listener during cleanup (#3438)
• vscode-ide-companion: preserve split stream message ordering (#3450)
• core: normalize Windows PATH for MCP stdio servers (#3451)
• core: prevent malformed permission rules from becoming tool-wide catch-alls (#3467)
• cli: pin /recap above input and align defaults with fastModel (#3478)
• cli: rework session recap rendering and add blur threshold setting (#3482)
• mcp: make the OAuth authorization URL clickable when wrapped (#3489)
• core: recover from truncated tool calls via multi-turn continuation (#3313)
• editor: detect Zed.app on macOS when CLI is not in PATH (#3303)
• openai: when samplingParams is set, pass it through verbatim (#3458)
• Handle missing xdg-open (ENOENT) gracefully to prevent crash (#1675)
• core: use empty string instead of null for reasoning-only assistant content (#3499)
• cli: inject plan/subagent/arena system reminders in ACP (#1151) (#3479)
• core: reject truncated subagent write_file calls (#3505)

Performance

• vscode: fix input lag in long conversations (#2550)

Documentation

• fix Windows install command to work in both CMD and PowerShell (#3252)
• update authentication methods to reflect OAuth discontinuation (#3325)

Other

• test(core): stabilize glob truncation tests (#3322)
• test(integration): match new cron notification format in interactive tests (#3402)
• Fix typo in class name (#2189)
• test(core): update scheduler registry mock (#3415)
• ci(stale): enable 60+30 stale/close policy for pull requests (#3375)
• Revert "feat(core): add dynamic swarm worker tool" (#3468)
• test(integration): switch settings-migration probe from --help to mcp list (#3486)

0.14.5 - 2026-04-15

Added

• cli/sdk: expose /context usage data in non-interactive mode and SDK API (#2916)
• cli: add startup performance profiler (#3232)
• core: implement fork subagent for context sharing (#2936)
• vscode-ide-companion: add /account for account display (#2984)
• acp: LLM-based message rewrite middleware with custom prompts (#3191)
• auth: discontinue Qwen OAuth free tier (2026-04-15 cutoff) (#3291)

Fixed

• core: detect rate-limit errors from streamed SSE frames (#3246)
• vscode: limit session tab title length to prevent tab bar overflow (#3249)
• core: respect custom Gemini baseUrl from modelProviders (#3212)
• core: allow thought-only responses in GeminiChat stream validation (#3251)
• cli: make /bug easier to open in terminals without hyperlink support (#3257)
• cli: ignore literal Tab input in BaseTextInput (#3270)
• channels/dingtalk: prioritize senderStaffId over senderId for allowedUsers matching (#3294)
• cli: block discontinued qwen-oauth model selection in ModelDialog (#3299)

0.14.4 - 2026-04-13

Added

• cli: CJK word segmentation and Ctrl+arrow navigation optimization (#2942)
• replace text input with model picker for Fast Model in /settings (#3120)
• show description for active setting in /settings dialog (#3116)
• i18n: add French (fr-FR) locale support (#3126)
• cli: queue input editing — pop queued messages for editing via ↑/ESC (#2871)
• channels: add voice message support in TelegramAdapter (#3150)
• cli: support tools.sandboxImage in settings (#3146)
• cli: warn when workspace overrides global modelProviders (#3148)
• hooks: Add StopFailure and PostCompact hook events (#2825)
• core: intelligent tool parallelism with Kind-based batching and shell read-only detection (#2864)
• add contextual tips system with post-response context awareness (#2904)
• subagents: propagate approval mode to sub-agents (#3066)
• skills: add model override support via skill frontmatter (#2949)
• cli: support bare exit/quit commands to exit the CLI (#3201)
• subagents: add disallowedTools field to agent definitions (#3064)
• core: add microcompaction for idle context cleanup (#3006)

Changed

• merge test-utils package into core (#3200)

Fixed

• vscode: force fresh ACP session on new-session action (#2874)
• cli: prioritize slash command completions (#3104)
• cli: improve markdown table rendering in terminal (#2914)
• prevent statusline script from corrupting settings.json (#3091)
• cli: check NEWLINE before SUBMIT in TextInput multiline mode (#3094)
• input: preserve tab characters in pasted content (#3045)
• use latest assistant token count on resume instead of stale compression checkpoint (#3109)
• upgrade normalize-package-data to 7.0.1 (fixes DEP0169 warning) (#2865)
• core: cap recursive file crawler at 100k entries to prevent OOM (#3138)
• channels: apply proxy settings to channel start command (#3136)
• lazy-load channel plugins to eliminate DEP0040 startup warning (#3134)
• core: fall back to CLI confirmation when IDE diff open fails (#3031)
• core: handle empty OAuth refresh response body (#3123)
• followup: fix follow-up suggestions not working on OpenAI-compatible providers (#3151)
• cli: recover from stuck bracketed-paste mode and keep Ctrl+C reachable (#3181)
• cli: set qwen3.5-plus as default model for Coding Plan (#3193)
• core: respect respectGitIgnore setting in @file injection path (#3197)
• core: show clear error when MCP server cwd does not exist (#3192)
• cli: honor --openai-api-key in non-interactive auth validation (#3187)
• cli: stop refilling input with prior prompt on cancel (#3208)
• core: allow Unicode characters in agent names (#3194)

Documentation

• readme: Add announcement for Qwen OAuth free tier policy adjustment (#3207)
• update quota exceeded alternatives to OpenRouter and Fireworks (#3217)

Other

• chore: remove legacy directories (.gcp, .aoneci, hello, .allstar) (#3199)
• ci(release): parallelize release validation (#3132)
• chore: bump version to 0.14.4 (#3209)

0.14.3 - 2026-04-10

Added

• plan: add "Yes, restore previous mode" option when exiting plan mode (#3008)
• review: enhance /review with deterministic analysis, autofix, and security hardening (#2932)
• ui: add customizable status line with /statusline command (#2923)

Changed

• centralize IDE diff interaction in CoreToolScheduler (#2728)
• rename verboseMode to compactMode for better UX clarity (#3075)

Fixed

• ui: Remove dead dirs state and unused hook parameter from InputPrompt (#2891)
• followup: prevent tool call UI leak and Enter accept buffer race (#2872)
• core: add getDefaultPermission and allowExternalPaths to ripGrep tool (#2948)
• webui: fix chat input scrollbar not draggable in VS Code plugin (#3038)
• bundle: inline tree-sitter WASM for bundled installs (#2985)
• cli: serialize subagent confirmation focus to prevent concurrent input conflicts (#2930)
• permissions: match env-prefixed shell commands against saved permission rules (#2850)
• prevent Shift+Tab from accepting prompt placeholder suggestion (#3060)
• weixin: add missing iLink headers to QR code login flow (#3044)
• improve /model --fast description clarity (#3077)
• cli: add 'detail' subcommand to /context command (#3042)
• persist ProceedAlways permission outcome in compact mode (#3069)
• add --fast hint to /model description for discoverability (#3086)

Other

• chore: remove outdated pr-review skill (#3028)
• test: add tests for confirmation-bus, prompt-registry, and cli/core modules (#2272)
• [codex] fix checkpointing init in non-repo directories (#3041)
• chore: bump version to 0.14.3 (#3112)

0.14.2 - 2026-04-08

Added

• cli: implement /plan command for plan mode (#2921)
• core: thinking block cross-turn retention with idle cleanup (#2897)
• core: adaptive output token escalation (8K default + 64K retry) (#2898)
• add bugfix workflow, test-engineer agent, and debugging skills (#2881)
• add qwen3.6-plus model to ModelStudio Coding Plan (#3015)

Fixed

• vscode-ide-companion: fix blank screen in VS Code 0.14.1 webview (#2959)
• hooks: preserve null exit code from signal kills instead of collapsing to 0 (#2976)
• cli: disable follow-up suggestions by default (#2954)
• cli: fix csiUPrefix error in Linux/Wayland (#2995)
• cli: sync packages/cli version and sandboxImageUri to 0.14.2 (#3026)

Other

• bump version to 0.14.2 (#3020)

0.14.1 - 2026-04-07

Added

• cli: enhance /btw side question with improved prompt and Ctrl+C/D cancel (#2776)
• cli, webui: add follow-up suggestions feature (#2525)
• webui: unify remaining tool display labels (#2595)
• allow Ctrl+Y to skip rate-limit retry delay immediately (#2420)
• prompt: add dangerous actions behavior guidance in system prompt (#2889)
• core: implement mid-turn queue drain for agent execution (#2854)
• to #2767, support verbose and compact mode swither with ctrl-o (#2770)

Changed

• tools: remove duplicate proxy setup in WebFetchTool (#2888)

Fixed

• hooks: clean up abort listener in error handler (#2841)
• cli: commit pending AI response before adding hook system message (#2848)
• subagents: preserve session subagents during cache refresh (#2895)
• telegram: send only failed chunk as plaintext fallback (#2894)
• auth: only release token refresh lock if it was acquired (#2893)
• extensions: handle individual extension update check failures (#2892)
• mcp: clear OAuth callback timeout on all completion paths (#2890)
• mcp: clean up directory listener on connect failure (#2896)
• permissions: allow non-core tools to bypass coreTools allowlist (#2843)
• prevent output-language.md from being overwritten on startup (#2842)
• cli: restore ? shortcuts in vim normal mode (#2884)
• cli: prevent ideCommand failure from breaking all slash commands… (#2822)
• improve ACP connection reliability with spawn retry and auto-reconnect (#2804)
• vscode: inherit model selection for new chat tabs (#2802)
• hooks: parse JSON output on exit code 2 to preserve hook additionalContext (#2815)
• cli: remove quote-based drag detection to prevent input lag (#2837)
• cli: restore previous theme on /theme cancel (refs #2833) (#2834)
• extensions: await async calls in extension refresh chain (#2835)
• cli: preserve runtime-added models when saving settings (#2455)
• tools: exit_plan_mode now exits correctly in YOLO mode (#2586)
• vscode: remove @vscode/vsce from devDependencies to fix local build (#2824)
• webui: remove @qwen-code/qwen-code-core dependency (#2902)
• core: coerce stringified JSON values for anyOf/oneOf MCP tool schemas (#2858)
• weixin: add missing iLink-App-Id and iLink-App-ClientVersion headers (#2943)

Other

• chore: bump version to 0.14.1 (#2849)
• Fix Markdown table cell separator escaping in MarkdownDisplay.tsx (#2463)
• Remove CODEOWNERS file (#2937)

0.14.0 - 2026-04-03

Added

• hooks: remove experimental flag and add disabled state UI (#2781)
• vscode: add retry logic and auto-reconnect for ACP connection (#2666)
• add cross-provider model selection for subagents (#2698)
• extension: Add npm registry support for extension installation (#2719)
• cron: add in-session loop scheduling with cron tools (#2731)
• channels: add extensible Channels platform with plugin system and Telegram/WeChat/DingTalk channels (#2628)
• mcp: add reconnect command and implement auto-reconnect logic (#2428)

Changed

• ui: improve hook event handling with dedicated history items (#2696)
• PR #2666 ACP retry/reconnect logic (#2792)

Fixed

• add .qwen path replacement in markdown files during extension install (#2769)
• normalize proxy URLs to support addresses without protocol prefix (#2745)
• make /compress handle tool-heavy conversations correctly (#2659)
• core: robustly resolve tree-sitter WASM path for symlinked CLI installations (#2764)
• prevent subagent telemetry from overwriting main agent footer context (#2765)
• upgrade @lydell/node-pty to 1.2.0-beta.10 to fix PTY FD leak on macOS (#2777)
• allow web fetch approvals in plan mode (#2763)
• prevent orphan ACP processes on tab close and clean up MCP subprocesses on shutdown (#2662)
• cli: enhance KeypressProvider with kitty sequence timeout manage… (#2612)
• delete design doc (#2789)
• resolve punycode to userland package and skip env var test in sandbox (#2796)
• hide skills with cron allowedTools when cron is disabled (#2811)

Other

• Enhance /review: add verification, false positive control, and PR comments (#2687)
• chore(channels): make plugin-example private and remove from release workflow (#2801)
• 🎉 feat: add Qwen3.6-Plus model support (#2820)

0.13.2 - 2026-03-30

Added

• add bundled qc-helper skill, qwen-code-claw reference, and README claw guide (#2623)

Fixed

• docs: update references from Bailian to ModelStudio in README an… (#2714)
• shell: resolve Git Bash path for node-pty on Windows (#2733)
• resolve /clear command and ESC key lag caused by hooks system (#2656)
• preserve original line endings (CRLF/LF) when editing files (#2707)
• core: resolve tree-sitter wasm path for symlinked CLI (#2744)
• cli: prevent terminal response leakage on high-latency SSH (#2718)
• shell: remove command substitution deny check from getDefaultPermission (#2747)
• make list_directory integration test more deterministic (#2752)

Documentation

• clarify envKey usage and add env field examples (#2715)

Other

• chore: bump version to 0.13.1 (#2716)
• chore: release v0.13.2 (#2750)

0.13.1 - 2026-03-27

Added

• hooks: Add comprehensive hook execution telemetry (#2421)
• hooks ui: refactor ui for Qwen Code hooks (#2602)
• human-readable permission labels, deny rule feedback, and multi-dir search improvements (#2637)
• auth: implement Alibaba Cloud Standard API Key support (#2668)

Fixed

• extensions: support non-GitHub git URLs for extension installation (#2539)
• cli: /memory show --project and --global now display all configured context files (#2368)
• mcp: restore trust+isTrustedFolder permission check in getDefaultPermission (#2642)
• cli: preserve selected auth type on startup auth failure (#2080)
• vscode-ide-companion: improve ACP error handling to prevent silent loading hangs (#2546)
• vscode-ide-companion: silence secondary sidebar warning on older VS Code versions (#2545)
• lsp: improve C++/Java/Python language server support (#2547)
• vscode-ide-companion: preserve model metadata on switch (#2591)
• windows: support git bash/MSYS2 shell detection on Windows (#2645)
• shell: handle PTY race condition errors gracefully (#2611)
• acp-integration/agent: clear stale subagent diff confirmation after IDE accept (#2631)
• use config working directory for OpenAI logger path resolution in ACP mode (#2675)
• @ file search stops working after selecting a slash command (#2694)
• acp: align permission flow across clients (#2690)

Documentation

• add hooks documentation and fix JSON schema (#2679)

Other

• test(sdk): improve tool control docs and add pattern matching tests (#2644)
• test(sdk): improve permission message pattern matching (#2712)

0.13.0 - 2026-03-23

Added

• add system prompt customization options in SDK and CLI (#2400)
• hooks: implement hooks extension mechanism (#2352)
• core: execute task tools concurrently for improved performance (#2434)
• arena: Add agent collaboration arena with multi-model competitive execution (#1912)
• ui: Display token usage in the loading/progress indicator (#2445)
• vscode-ide-companion: add Tab key fill-only behavior for completions (#2431)
• add /context command to display context window token usage breakdown (#1835)
• support skills in .agents directory and other provider directories (#2202)
• add auth CLI command and Qwen Code Claw skill (#2440)
• export: add metadata and statistics tracking (#2328)
• hooks: Implement 10 core event hooks for session lifecycle and tool execution (#2203)
• support permission (#2283)
• add .agents/skills as a skill provider directory (#2476)
• vscode-ide-companion: add image paste support (#1978)
• storage: support configurable runtime output directory (#2127)
• core: add Explore agent and rename TaskTool to AgentTool (#2489)
• hooks: use extension dir files instead of tmp dir files (#2478)
• cli: add /btw slash command for ephemeral side questions (#2371)

Changed

• core: improve error handling and quota detection (#2458)
• Refactors the VS Code file completion system to use fuzzy search (#2437)

Fixed

• pipeline: handle duplicate finish_reason chunks from OpenRouter (#2403)
• cli: show newest-first history for Ctrl+R command search (#2425)
• Ensure message_start and message_stop events are paired in SDK streaming (#2448)
• core: add truncation support for MCP tool output (#2446)
• vscode-ide-companion: update URI handling for Windows paths (#2457)
• test: update LoadingIndicator snapshot for correct output alignment (#2469)
• correct token limits for MiniMax-M2.5 and GLM models (#2470)
• update TOS link in VS Code extension README (#2495)
• preserve modalities during OpenAI logging request conversion (#2473)
• clean up ACP connection state when child process exits (#2472)
• vscode-ide-companion: pass proxy configuration to CLI (#2501)
• include bundled skills directory in published package (#2521)
• update Discord invite link to permanent URL (#2535)
• web-fetch: add simplified system instruction to prevent AI greeting responses (#2610)
• hooks: terminate hook child processes when user exits CLI (#2607)

Documentation

• rename QWEN.md to AGENTS.md to follow community best practices (#2527)
• add Screenshots/Video Demo section to PR template (#2533)

Other

• chore: bump version to 0.13.0 (#2451)
• Fix shell permission parsing and test-created debug artifacts (#2536)

0.12.6 - 2026-03-17

Fixed

• improve max_tokens handling with conservative defaults (#2438)

Other

• chore: bump version to 0.12.6 (#2442)

0.12.5 - 2026-03-16

Fixed

• shell: resolve Windows encoding issues for non-ASCII output (#2423)

Other

• test(sdk): simplify integration tests for reliability (#2410)
• chore: bump version to 0.12.5 (#2422)

0.12.4 - 2026-03-16

Added

• skills: add bundled /review skill for out-of-the-box code review (#2348)
• skills: add docs audit and update helpers (#2397)

Fixed

• insight: handle individual LLM failures in qualitative insights (#2341) (#2361)
• core: add deepseek-r1 to output token limit patterns (#2362)
• i18n: localize slash command descriptions (#2333)
• core: guard against empty choices in convertOpenAIResponseToGemini (#2364)
• extension: disable symlinks on Windows during git clone to fix install failure (#2286)
• core: reject PDF files to prevent session corruption (fixes #2020) (#2024)
• cli: allow /dev/ptmx and /dev/ttys* in macOS permissive sandbox (#2391)
• correct hooks JSON schema type definition (#2280)
• core: strip orphaned user entries before retry to prevent API errors (#2367)
• core: correctly capture rapid pty outputs in interactive shell mode (#2389)
• vscode: prevent race conditions in prompt cancellation and streaming (#2374)
• core: improve shell tool truncation, simplify tool output handling, and remove summarization (#2388)
• remove redundant plan files (#2407)
• core: normalize Windows PATH-like env keys for shell execution (#1904)
• auto-detect max_tokens from model when not set by provider (#2356)

Documentation

• explain Docker sandbox runtime and Java usage (#1642)
• integration: add ACP Registry for Zed and JetBrains integration docs (#2372)

Other

• Docs/subagent system prompt limits (#2001)
• Keep rejected plan content visible in plan mode (#2157)
• chore(CODEOWNERS): remove required reviewers for vscode-ide-companion and webui packages (#2408)
• Increase DEFAULT_OUTPUT_TOKEN_LIMIT from 8K to 16K (#2411)

0.12.3 - 2026-03-13

Added

• mcp: improve OAuth auth UX - post-auth feedback, i18n, clear auth, and bug fixes (#2327)

Fixed

• ide: resolve IDE connection issues in some VSCode clients and optimize connection config lookup (#2322)
• core: correct GPT-5.x input token limit to 272K (#2345)
• shell: pass args as string on Windows to prevent quoting issues (#2347)
• core: disable node-pty on older Windows builds with broken ConPTY (#2349)
• improve qwen mcp add option handling for arrays (#2245)
• cli: prevent Ctrl+F from leaking to PTY as ^F artifact (#2350)
• core: remove duplicate exports in packages/core/src/index.ts (#2265)
• cli: remove unused debug log session setup in loadSettings (#2355)

Other

• Refactors FileSystemService interface to use ACP-aligned request/response objects (#2344)

0.12.2 - 2026-03-12

Added

• core: add truncation support to LS tool (#2324)

Fixed

• export command should use current session ID instead of loadLastSession (#2268)
• webui: add Tab key support to CompletionMenu (#2308)
• core: convert array content to string for DeepSeek API (#2320)
• improve ACP file operation error handling (#2298)
• remove QR code from OAuth authentication UI to prevent screen flickering (#2315)
• clear retry error messages promptly after auto-retry succeeds (#2326)

Other

• chore: add yiliang114 as code owner for vscode-ide-companion and webui (#2312)
• chore: Release v0.12.2 (#2307)

0.12.1 - 2026-03-11

Added

• cli: change temporary filename prefix to qwen-edit- (#2045)
• vscode-ide-companion: add sidebar view and multi-position chat layout (#2188)

Fixed

• mcp: use scopes from protected resource metadata (RFC 9728) (#2212)
• cli: clear static error message when starting new query (#2110)
• clean up MCP server display and add CONCAT merge strategy for mcp allowed/excluded lists (#2219)
• hooks: Fix failing hook integration tests by updating hook scripts to create hook_invoke_count.txt (#2230)
• hooks: Remove useless expect (#2238)
• core: skip openDiff in YOLO mode to prevent VS Code editor from opening (#2221)
• cli: suppress Windows pty resize race condition (#2289)
• vscode-ide-companion: map ENOENT errors to ACP RESOURCE_NOT_FOUND in readTextFile (#2291)

Other

• improve readability of context compression description (#2224)
• refactore: Start qwen after installation (#2290)

0.12.0 - 2026-03-09

Added

• add tabWidth support for code highlighting and replace tabs with spaces in CodeColorizer (#2077)
• export-html: viewer for tool call results (#2085)
• terminal-capture: add streaming capture with GIF generation (#2116)
• commands: add custom QC commands for GitHub workflows (#2117)
• add support for printable CSI-u keys in KeypressContext (#1827)
• add JSON Schema validation for VS Code settings (#1830)
• hooks: Implement hooks system infrastructure with CLI and UI management (#1988)
• shell: enable PTY by default and various enhancements (#2108)
• Enhance MCP Management TUI with dynamic enable/disable and runtime updates (#1831)
• Add interactive TUI for extension management (#2008)
• Implement AskUserQuestionTool for interactive user queries (#1828)

Changed

• cli: consolidate message components and fix leading icon display issues (#2120)
• unify sandbox configuration naming and improve telemetry config (#1793)
• acp: migrate ACP integration to @agentclientprotocol/sdk (#2063)

Fixed

• cli: parse markdown command frontmatter on Windows CRLF/BOM (#2078)
• cli: ignore stream-json input format in TTY mode to prevent hanging (#2047)
• core: prevent duplicate function-call yields from trailing stream chunks (#2125)
• ide: add async DNS check for host.docker.internal in container environments (#1817)
• handle symlinks during extension installation (#2056)
• preserve original encoding when reading/writing non-UTF-8 files (#2073)
• install: Add tips and fix installation issues for installation scripts (#2118)
• core: add independent retry budget for transient stream anomalies (#2126)
• windows: resolve silent failures caused by CRLF line endings (#1868) (#1890)
• cli: keep AGENTS.md enabled by default context reset (#2082)
• core: remove LLM-based loop detection and enable skipLoopDetection by default (#2092)
• keyboard: handle Kitty keypad private-use keycodes (#2137)
• hooks: fix result aggregator for userPromptSubmit and fix enable for integration test (#2139)
• hooks: Move enable from hooks to hookConfig and add max turns (#2156)
• Hooks online integration test failed (#2183)
• improve MCP Management & Extension Management TUI based on 0.12.0 feedback (#2208)
• test: use toContain instead of toBe for file content assertion (#2218)

Other

• chore: bump version to 0.12.0 (#2090)
• Refactor settings migration to sequential framework with atomic file writes (#2037)
• chore: add @DragonnZhang to CODEOWNERS (#2138)

0.11.1 - 2026-03-03

Added

• support AGENTS.md as default context file (#2018)
• cli: add Ctrl+Y shortcut to retry failed requests (#2011)
• cli: improve auth dialog UX with clearer three-option layout (#2030)
• i18n: strengthen output-language.md template to enforce language compliance (#2005)

Changed

• core: extract single tool-call execution path (#1999)

Fixed

• subagent: append output-language.md to subagent system prompt and prioritize project-level settings (#1993)
• core/rateLimit: add support for rate limit error code 1305 and custom retry error codes (#1995)
• logging: reduce excessive streaming output in session history logs (#2041)
• add modality defaults to prevent API errors when reading PDFs and other media (#1982)
• detect and protect against truncated tool call output (#2021)
• acp: add session/set_config_option method to enable config option updates from Zed UI (#2059)
• dashscope: support subdomain URL patterns for DashScope provider detection (#2060)

Documentation

• update installation instructions (#1994)

Other

• chore: bump version to 0.11.1 (#2026)
• Fix ACP protocol compatibility issues with Zed editor (#2017)

0.11.0 - 2026-02-28

Added

• Add clipboard image support and attachment UI to CLI (#1612)
• support MCP readOnlyHint annotation in plan mode (#1826) (#1837)
• Add insight command for personalized programming insights (#1593)
• auth: add automatic backup of settings.json before auth modification (#1952)
• cli: Increase /insight feature exposure via weighted tips (#2019)

Fixed

• Installation script permission check for arch os and add sudo check (#1877)
• normalize Windows paths to lowercase for case-insensitive session matching (#1768)
• enforce plan mode restrictions in ACP sessions (#1812)
• test: keep plan mode active during ACP integration test (#1956)
• change workspaceFolders capability to boolean for LSP servers (#1929)
• unblock input after ESC cancel (#1796)

Documentation

• enhance modelProviders documentation with comprehensive examples and behavior clarifications (#1927)
• fix documentation errors in commands and model-providers (#1962)

Other

• 📸 terminal-capture: CLI Terminal Screenshot Automation (#1840)
• chore: bump version to 0.11.0 (#1953)
• Merge coder-model and qwen3.5-plus, remove vision auto-switching (#1852)
• Rename GEMINI_CLI_INTEGRATION_TEST to QWEN_CODE_INTEGRATION_TEST and refactor sandbox user handling (#1966)

0.10.6 - 2026-02-24

Added

• add third-party models (glm-4.7, kimi-k2.5, qwen3-coder-next) to Coding Plan (#1907)
• runner: support auth_type for model configuration (#1874)
• update bailian coding plan models (#1931)

Fixed

• fs: Improve BOM detection with length check and codePointAt (#1857)
• update security vulnerability reporting channel (#1921)

Other

• chore: bump version to 0.10.5 (#1886)
• Fix release workflows: standardize notes generation and add prerelease labels (#1885)
• chore: exclude .qwen/commands/ and .qwen/skills/ from gitignore (#1847)

0.10.5 - 2026-02-18

Added

• add qwen3.5-plus model support for Coding Plan (#1867)

Other

• chore: bump version to 0.10.4 (#1864)

0.10.4 - 2026-02-18

Documentation

• add news banner about Qwen3.5-Plus launch (#1854)

Other

• Fix sandbox user permission in integration tests (#1843)
• Add Coding Plan Global/Intl region support (#1860)
• chore: bump version to 0.10.3 (#1863)

0.10.3 - 2026-02-16

Added

• update readme (#1853)

Documentation

• improve settings.json configuration guide with quick setup examples (#1850)

Other

• chore: bump version to 0.10.2 (#1844)

0.10.2 - 2026-02-14

Added

• add TPM throttling error handling with 1-minute retry delay (#1791)

Changed

• cli: unify Escape key handling in AppContainer (#1824)

Fixed

• Fix node installation permission issue in shell script (#1819)
• prevent AbortSignal listener memory leak (#1811)
• correct showLineNumbers default value to true (#1813)
• support JSON Schema draft-2020-12 for MCP tools (fixes #1818) (#1821)

Documentation

• update authentication documentation with Coding Plan setup guide (#1800)

Other

• chore: bump version to 0.10.1 (#1808)
• Add dev launch config and preserve existing NODE_OPTIONS (#1784)
• Fix abort listener accumulation in subagent while loop (#1825)
• Fix auth UI to use semantic theme colors and correct selection sync (#1823)
• Add --session-id support for CLI and SDK (#1822)

0.10.1 - 2026-02-11

Added

• add MCP tool progress update support in TUI and SDK mode (#1756)
• add Coding Plan authentication mode with unified AuthDialog (#1788)
• coding-plan: implement Coding Plan configuration management and update prompts (#1805)

Fixed

• Warning in installation shell script (#1771)
• ui: resolve model not updating in top-right corner (#1662)
• cli: use PowerShell Get-Command for Windows sandbox detection (#1604)
• prioritize local path detection in extension installation (#1770)
• auth-model-login-ui: prevent Enter key from triggering empty message submission (#1773)

Other

• Fix SDK MCP integration tests by updating hardcoded tool names to use constants (#1769)

0.10.0 - 2026-02-09

Added

• query: add support for resuming sessions with session ID (#1714)
• Remove Smart Edit tool and ClearcutLogger (#1684)
• sdk: add resume, continue options and extend authType support (#1726)
• debug mode output refactor — route console calls to logfile-first debugLogger (#1610)
• paste: add large paste placeholder and fix enter-submit on macOS (#1713)
• promote Agent Skills from experimental to stable (#1738)
• add source information tracking in telemetry logs (#1653)
• settings: add settings.env field for environment variable configuration (#1751)

Changed

• i18n: translate Agent as 智能体 (#1718)
• remove read_many_files tool, add readManyFiles utility for user @-commands (#1673)

Fixed

• docker: fix build error and enable manual version builds (#1722)
• settings: rename negative settings to positive naming (disable* -> enable*) (#1330)
• clarify is_background parameter is required in docs and examples (#1716)
• vscode-ide-companion: Fix UI display issues with server-side timestamp and file path extraction (#1682)
• ui: resolve auth not updating in top-right corner (#1670)
• use openai model instead of index=0 in acp integration test (#1733)
• cli: route sandbox diagnostic messages to stderr (#1735)
• cli: prevent Tab key from cycling approval mode when autocomplete is active on Windows (#1736)
• mcp: improve MCP server management and authentication (#1752)
• core: properly handle MCP multi-part tool results in OpenAI converter (#1755)
• integration-tests: correct MCP tool name in simple-mcp-server test (#1763)

Documentation

• Update Linux/Mac installation commands in README (#1739)

Other

• ci(sdk-release): use stable CLI tags for SDK releases (#1710)
• add hint for installing external source extensions (#1694)
• Feat/javasdk alpha 202501 (#1717)
• Add export command for session history with markdown and HTML formats (#1515)
• Add FORK_MODE support to ProcessTransport for Electron IPC integration (#1719)
• Fix ACP model selection to handle all configured authentication types (#1555)
• chore: Reduce Qwen OAuth free quota from 2000 to 1000 requests per day (#1730)
• Add CLI source selection for SDK releases and fix subagent output handler (#1732)
• Fix CLI argument parsing for /dist/cli/cli.js entry point (#1758)

0.9.1 - 2026-02-05

Added

• core: add symlink support for skill manager (#1690)
• Preserve UTF-8 BOM when editing files (#1680)

Fixed

• core: properly cleanup MCP server subprocesses on exit (#1285)
• cli: expand MCP @server: resource references (#1531)
• core: auto-enable WebFetch and WebSearch tools in Plan mode (#1686)
• normalize skill file content in extensions to handle BOM and CRLF (#1667)
• ci: honor manual preview version input (#1665)
• core: handle heredoc in command substitution guard (#1701)
• core: Preserve trailing whitespace in newString during edits (#1688)
• enable Shift+Tab shortcut in Windows PowerShell (#1607)
• core: enforce tool restrictions in subagents (#1691)

Other

• test(cli): stabilize AuthDialog ESC assertion (#1535)
• build: Improve build efficiency and add dev mode (#1681)
• [AnthropicContentGenerator] optimize: ADD cache_control for system and last user text message (#1613)

0.9.0 - 2026-02-03

Added

• core: improve error message when skill is invoked as tool (#1623)
• core: improve retry logic for better 429/5xx error handling (#1628)
• add extra_body support for OpenAI-compatible providers (#1654)
• add multi-modal input support (image, PDF, audio) across all content generators (#1564)
• clarify output formats for non-interactive mode (#1579)
• add concurrent runner for batch CLI execution (#1640)
• webui: implement unified UI architecture with shared component library (#1543)

Fixed

• Use resolved authType to initialize ACP agent (#1622)
• acp: stream subagent text + reasoning chunks (#1626)
• ensure output-language.md is created before config initialization (#1637)
• security: prevent command injection via newline bypass in shell command validation (#1638)
• React/React-DOM version inconsistency in package.json and lockfile (#1659)
• core: avoid passing undici agent to Anthropic SDK (#1663)
• vscode-ide-companion: fix race conditions and improve @ file completion search (#1676)

Other

• chore: bump version to 0.8.2 (#1632)
• Add parentToolCallId and subagentType for ACP subagent tracking (#1620)
• Fix Claude plugin resource collection to respect marketplace config (#1639)
• Support model selection through ACP in vscode ide companion (#1582)
• Add Zed extension for Qwen Code agent server (#1630)
• Add experimental LSP support for code intelligence (#1401)
• chore: bump version to 0.9.0 (#1661)
• Add contextWindowSize Configuration Support (#1539)

0.8.2 - 2026-01-30

See GitHub release for details.

0.8.1 - 2026-01-27

Added

• i18n: Add Japanese language support and fix menu labels in other languages (#1392)
• Add Portuguese (pt-BR) language support with complete translations and refactor i18n architecture for better language management. (#1616)
• add skills and agents display to extension list with i18n support (#1629)

Fixed

• replace EnvHttpProxyAgent with ProxyAgent to suppress experimental warning (#1624)

Other

• test: improve SDK integration test reliability with createResultWaiter and ProcessTransport error handling (#1627)
• chore: bump version to 0.8.1 (#1631)

0.8.0 - 2026-01-27

Added

• cli: use dim colors for YOLO/auto-accept mode borders (#1476)
• Redesign CLI welcome screen and settings dialog (#1513)
• extensions: add detail command and improve extension validation (#1580)
• add runtime-aware fetch options for Anthropic and OpenAI providers (#1516)
• extensions: add plugin selection UI for Claude marketplace (#1592)
• make DiffRenderer respect ui.showLineNumbers setting (#1561)
• Implement temporary dismissal for feedback dialogs with persistent prompting (#1590)

Fixed

• replace spawn shell option with explicit shell args to avoid Node.js DEP0190 warning (#1234)
• skip non-existent file imports instead of warning (ENOENT) (#1563)
• correct schema field name for context.loadFromIncludeDirectories (#1609)
• vscode-ide-companion: platform-specific builds with optimized VSIX packaging (#1586)
• cli: pass paths to read_many_files in ACP (#1614)
• Add toolName metadata for ACP tool call messages (#1615)
• cli input stream handling and error management (#1588)

Documentation

• add Trendshift badge to README (#1553)

Other

• chore: remove tiktoken dependency and use API-reported token counts (#1526)
• Add /bug command to non-interactive mode (#1552)
• Feat/extension (#1534)
• fix dependences of core pkg (#1574)
• fix github pkg dependence (#1576)
• fix prompts denpendence (#1578)
• Add VSCode IDE Companion Release Workflow (#1542)
• Update command usage in add.ts to reflect new name (#1572)
• Security: Fix awk/sed Command Injection in READ_ONLY_ROOT_COMMANDS (#1601)
• Simplify permission response handling and fix edit failure and VSCode diff issues (#1581)

0.7.2 - 2026-01-20

Added

• cli: add settings support for experimental skills (#1497)
• Improve QWEN. md file loading by filtering system files and limiting scope (#1486)
• add user feedback dialog (#1465)

Fixed

• include --acp flag in tool exclusion check (#1499)
• vscode-ide-companion: simplify ELECTRON_RUN_AS_NODE detection and improve README (#1496)
• mistranslation of token (#1508)
• unable to remove MCP server when only one element exists (#1490)
• core: parse skills frontmatter with CRLF/BOM (#1528)
• cli: relocate skills setting to experimental namespace (#1538)
• acp: implement session/set_model method for JetBrains compatibility (#1521)
• resolve arrow key navigation conflict between history and completion (#1519)
• cli: isolate modelConfigUtils tests from system env vars (#1545)
• acp: propagate ENOENT errors correctly and centralize error codes (#1550)
• Update Qwen OAuth model information (#1548)

Documentation

• auth: add Coding Plan documentation (#1509)

Other

• Fix credential management and authentication flows with improved generation config preservation (#1510)

0.7.1 - 2026-01-14

Fixed

• docs (#1485)

Other

• Reduce slow quit by trimming skills watchers (#1489)
• Fix timing issue in LoggingContentGenerator initialization (#1492)
• chore: bump version to 0.7.1 (#1494)

0.7.0 - 2026-01-14

Added

• Modify the selection order of user Settings and workspace Settings (#1433)
• multi-provider models config support (#1291)
• skills: add experimental /skills command + hot reload (#1436)
• shell: add optional timeout for foreground commands (#1469)
• Customizing the sandbox environment (#1473)

Changed

• convert IDE context from JSON to plain text format (#1424)

Fixed

• core: ensure OAuth URL always displayed in headless mode (#1426)
• multi provider cold start issue (#1439)
• cli: /memory show respects context.fileName (#1428)
• resolve external editor launch failure on macOS and Windows (#1351)
• core: handle missing delta in OpenAI stream chunks (#1448)
• cli: default sandbox UID/GID mapping on Linux (#1453)
• shell: prevent console window flash on Windows for foreground tasks (#1464)
• cli: warn on deprecated/unknown settings keys (#1427)
• core: improve OAuth fetch-failed diagnostics (#1457)
• SDK release workflow and stability improvements (#1462)
• vscode-ide-companion: Fix cross-platform CLI terminal execution (#1474)
• cli: improve error message display for object errors (#1386)
• Improve qwen-oauth fallback message display (#1480)
• docs errors and add community contacts (#1484)

Documentation

• vscode-ide-companion: update vscode extension readme (#1472)
• add integration guide for JetBrains IDEs (#1411)

Other

• chore: bump version to 0.7.0 (#1434)
• Support Jupyter Notebook (.ipynb) File Code Selection (#1460)
• Feature/add custom headers support (#1447)
• Fix auth type switching and model persistence issues (#1478)
• Skip flaky permission control test (#1482)

0.6.2 - 2026-01-12

See GitHub release for details.

0.6.1 - 2026-01-07

Added

• i18n: auto-detect LLM output language from system locale (#1247)
• i18n: update Russian translation with new strings (#1293)
• i18n: add German language support (#1378)
• graduate --experimental-acp to stable --acp flag (#1355)
• cli: add direct argument support for /approval-mode command (#1391)
• Optimize the issue where an error message indicating unfriendli… (#1282)

Fixed

• core: coerce string boolean values in schema validation (#1284)
• cli: skip update check when disableUpdateNag is true (#1397)
• improve tool execution feedback in non-interactive mode (#1383)
• exit with non-zero code on API errors in text mode (#1376)
• preserve whitespace in thinking content for stream-json output format (#1365)
• improve windows background process handling and cleanup (#1146)
• cli,core: honor tools.core / tools.allowed in non-interactive runs (#1406)
• core: don’t force reasoning/topP defaults for OpenAI-compatible APIs (#1415)

Documentation

• add AionUi to ecosystem section (#1360)

Other

• Fix multi-language and documentation related issues. (#1332)
• support merge ChatCompletionContentPart && add filterEmptyMessages (#1288)
• Feat/javasdk (#1412)
• Doc/qwencode java (#1414)
• Fix resume command broken after new chat (#1374)
• chore: bump version to 0.6.1 (#1423)
• [OpenaiContentGenerate] convertOpenAIResponseToGemini record thoughtsTokenCount (#1393)

0.6.0 - 2025-12-26

Added

• add a link to Gemini CLI Desktop for Qwen Code users who prefer desktop UIs (#286)
• add Anthropic provider, normalize auth/env config, and centralize logging (#1331)
• vscode-ide-companion: in/output part in the bash toolcall can be clicked to open a temporary file (#1345)
• support /compress and /summary commands for non-interactive & ACP (#1322)

Fixed

• cli path parsing issue in Windows (#1321)
• mcp: update OAuth client name for Figma MCP server compatibility (#1302)

Documentation

• readme: clarify value props, usage modes (#1312)

Other

• Add Gemini provider, remove legacy Google OAuth, and tune generation … (#1297)
• Add experimental Skills feature (#1314)
• chore: revert sdk-typescript version to 0.1.0 and update release workflow (#1325)
• Follow up on pr #1331 (#1340)
• fix one flaky integration test (#1343)
• Enhance VS Code extension description with download link (#1341)
• fix one flaky integration test (#1349)
• chore: improve release-sdk workflow (#1334)
• context left on vscode ide companion (#1327)

0.5.2 - 2025-12-22

Other

• pump version to 0.6.0 (#1309)
• Improve robustness of getProcessInfo with try-catch and empty output fallback (#1310)
• fix e2e workflow (#1311)

0.5.1 - 2025-12-19

Added

• expose gitCoAuthor setting in settings.json and document it (#1228)
• ui: add /resume slash command to switch between sessions (#1239)

Fixed

• handle case-insensitive path comparison in glob tool on Windows (#1174)
• ide: rename Gemini references to Qwen and fix IDE connection path (#1205)
• add configurable OpenAPI 3.0 schema compliance for Gemini compatibility (#1186) (#1214)
• cli: handle PAT tokens and credentials in git remote URL parsing (#1225)
• cli: add -r and -C aliases for --resume and --continue options (#1286)
• default values of sampling params (#1269)
• vscode-ide-companion: Optimize stream termination handling and fix style layering issues (#1261)
• optimize windows process tree retrieval to prevent hang (#1231)

Documentation

• add comprehensive MCP Quick Start guides and examples (#796)
• restructure docs to follow the Claude Code organization (#1260)

Other

• Add chat recording toggle (CLI + settings) and disable recording in tests (#1254)
• pump version to 0.5.1 (#1259)
• remove one flaky integration test (#1275)
• docs：Fix the errors in the document (#1266)
• Bundle CLI into SDK package and separate CLI & SDK E2E tests (#1265)
• chore(vscode-ide-companion): update vscode engine version from ^1.99.0 to ^1.85.0 (#1262)
• IDE companion discovery: switch to ~/.qwen/ide lock files (#1257)

0.5.0 - 2025-12-13

Added

• i18n: add Russian language support (#1238)
• show session resume command on exit (#1219)
• add terminal bell setting to enable/disable audio notifications (#1194)

Changed

• vscode-ide-companion: optimize CLI detection and version management (#1248)

Fixed

• remove redundant if-check and add tests for OpenAI converter (#1235)
• vscode-ide-companion: improve cross-platform compatibility in prepackage script (#1249)

Other

• test(cli): add tests for /language command and fix LLM output language parsing (#1236)
• Add ACP authenticate update message (#1240)
• Remove obsolete “corgi mode” (#1245)
• Fix/vscode ide companion completion menu content (#1243)
• Bundle CLI into VSCode release package (#1246)

0.4.1 - 2025-12-12

Added

• ui: remove vertical borders from input prompt for easier copy/paste (#1191)
• VSCode Extension Implementation (#1059)
• update references from Gemini to Qwen in setup commands and gitignore handling (#1156)
• Add channel field support for client identification (#1226)

Fixed

• prefer UTF-8 encoding for shell output on Windows when detected (#1157)
• update vulnerable dependencies (glob, jws, tar, js-yaml) (#1189)
• 修复在docker环境中无法连接ide的问题 (#1230)
• vscode-ide-companion/auth: deduplicate concurrent authentication calls (#1223)

Other

• pump versionm to 0.4.1 (#1177)
• Feat/acp usage metadata (#1176)
• pump version to 0.5.0 (#1233)

0.4.0 - 2025-12-06

Added

• basic TypeScript SDK (#1103)

Fixed

• shell-utils: resolve command detection on Ubuntu by using shell for builtins (#1123)
• update timeout settings and default logging level in SDK (#1165)

Other

• Session-Level Conversation History Management (#1113)
• pump version to 0.4.0 (#1132)
• skip one flaky integration test (#1137)
• Skip acp integration test in sandbox environment (#1141)
• test: skip qwen-oauth test in containerized environments (#1150)
• Remove /quit-confirm flow (#1148)
• DeepSeek V3.2 Thinking Mode Integration (#1134)
• Custom tools support via SDK controlled MCP servers (#1147)
• test: separating integration tests for the CLI and SDK (#1161)
• test: skip unstable e2e test (#1166)

0.3.0 - 2025-11-28

Added

• i18n: Add Internationalization Support for UI and LLM Output (#1058)

Fixed

• ci: remove non-existent label from release failure issue creation (#1097)
• reset authType settings (#1091)

Other

• Headless enhancement: add stream-json as input-format/output-format to support programmatically use (#926)
• chore: pump version to 0.3.0 (#1085)
• Improve Usage Statistics by Moving Key Snapshot Fields into Properties (#1090)

0.2.3 - 2025-11-20

Changed

• auth: enhance useAuthCommand to include history management … (#1077)

Fixed

• character encoding corruption when executing the /copy command on Windows. (#1069)
• remove broken link (#1074)

Other

• chore: pump version to 0.2.3 (#1073)
• Disable Prompt Completion Feature (#1076)
• Replace spawn with execFile for memory-safe command execution (#1068)

0.2.2 - 2025-11-19

Added

• openApi configurable window (#1019)
• add support for alternative cached_tokens format in OpenAI conv… (#1035)
• add support for Trae editor (#1037)

Changed

• auth: save authType after successfully authenticated (#1036)

Fixed

• core: add modelscope provider to handle stream_options (#848)
• Improve ripgrep binary detection and cross-platform compatibility (#1060)
• skip problematic integration test (#1065)

Other

• chore: pump version to 0.2.2 (#1027)
• 🎯 Enhance QwenLogger with OS Platform and Version Metadata (#1053)
• Add Terminal Attention Notifications for User Alerts (#1052)
• Add (limited) slash command support for ACP integration. (#1020)
• Fix integration tests (#1062)

0.2.1 - 2025-11-13

Added

• enhance zed integration with TodoWriteTool and TaskTool support (#992)

Fixed

• Stream parsing for Windows Zed integration (#996)
• print request errors for logging only in debug mode (#1006)

Other

• chore: pump version to 0.2.1 (#1005)
• 🔧 Refactor: Standardize Tool Naming and Configuration System (#1004)
• Fix incorrect tools list format in subagent template documentation (#1026)
• 🎯 PR: Improve Edit Tool Reliability with Fuzzy Matching Pipeline (#1025)
• Add Interactive Approval Mode Dialog (#1012)
• Change deepseek token limits regex patterns for deepseek-chat (#817)

0.2.0 - 2025-11-07

Added

• Simplify and Improve Search Tools (glob, grep, ripgrep) (#969)

Changed

• Unifying the system information display between /about and /bug commands (#977)

Fixed

• VSCode detection null check and debug message optimization (#983)

Other

• chore: pump version to 0.1.5 (#974)
• 🎯 Feature: Customizable Model Training and Tool Output Management (#981)
• chore: pump version to 0.2.0 (#991)

0.1.5 - 2025-11-07

Added

• Simplify and Improve Search Tools (glob, grep, ripgrep) (#969)

Changed

• Unifying the system information display between /about and /bug commands (#977)

Fixed

• VSCode detection null check and debug message optimization (#983)

Other

• chore: pump version to 0.1.5 (#974)
• 🎯 Feature: Customizable Model Training and Tool Output Management (#981)
• chore: pump version to 0.2.0 (#991)

0.1.4 - 2025-11-05

Added

• support for custom OpenAI logging directory configuration (#972)

Fixed

• handle AbortError gracefully when loading commands (#936)

Other

• chore: pump version to 0.1.4 (#962)
• chore: Web Search Tool Refactoring with Multi-Provider Support (#885)
• Fix kimi2 token limits (#970)

0.1.3 - 2025-11-04

Fixed

• Include macOS Seatbelt Sandbox Files in NPM Package (#949)

Other

• chore: pump version to 0.1.3 (#939)
• 🐛 Fix: /ide install command fails on Windows (#957)
• Fix unhandled promise rejection on connecting to VSCode companion (#958)

0.1.2 - 2025-10-31

Fixed

• Use runtime session ID in /bug command (#927)
• update tool name from Gemini to Qwen Code in ToolsList component… (#933)
• settings: add version field to prevent partial migration corruption (#937)

Other

• chore: pump version to v0.1.2 (#907)
• fixbug: fix qwen help des (#915)
• 🔍 Refactor and Enhance Ripgrep Tool (#930)
• change Launch Gemini CLI to Qwen Code CLI in help information (#929)
• Fix Chat Compression System Instruction and Empty Summary Edge Case (#935)

0.1.1 - 2025-10-29

Fixed

• e2e test (#905)

Other

• chore: pump version to 0.1.1 (#883)
• fix input filter (#892)
• 🐛 Bug Fixes Release v0.1.1 (#898)
• [to #12345678] docs: update excludeTools documentation in extensions … (#904)

0.1.0 - 2025-10-27

Fixed

• Invalid Tool Calls Due to Improper Request Cancellation (#790)
• remove unavailable options (#685)
• token limits for qwen3-max (#724)
• add missing trace info and cancellation events (#791)
• unable to quit when auth dialog is opened (#804)

Documentation

• add /model command documentation (#872)

Other

• chore: remove default topp & temperature value (#785)
• Fix and update the token limits handling (#754)
• chore: re-organize labels for better triage results (#819)
• Sync upstream Gemini-CLI v0.8.2 (#838)
• chore: Adjusted docs directory structure (#864)
• 📦 Release qwen-code CLI as a Standalone Bundled Package (#866)
• Standardize Tool Output Format for Better LLM Communication (#881)

0.0.14 - 2025-09-29

Added

• Implement Plan Mode for Safe Code Planning (#658)
• Add Qwen3-VL-Plus token limits (256K input, 32K output) (#720)

Fixed

• TaskTool Dynamic Updates (#697)

Other

• chore: bump version to 0.0.13 (#695)
• 🐛 Remove unreliable editCorrector that injects extra escape characters (#713)
• Fix/qwen3 vl plus highres (#721)
• 🚀 feat: DashScope cache control enhancement (#735)

0.0.13 - 2025-09-24

Added

• add OpenAI and Qwen OAuth auth support to Zed ACP integration (#678)
• add yolo mode support to auto vision model switch (#652)

Fixed

• output token limit for qwen (#664)
• auth hang when select qwen-oauth in Zed (#684)
• ripgrep load issue (#676)

Other

• chore: bump version to 0.0.12 (#662)
• 🐛 Fix: Resolve Markdown list display issues on Windows (#693)

0.0.12 - 2025-09-19

Added

• Enhance /init command with confirmation prompt (#624)

Fixed

• Windows Multi-line Paste Handling with Debounced Data Processing (#627)
• subagent system improvements and UI fixes (#638)
• reset is_background (#644)
• switch system prompt to avoid malformed tool_calls (#650)
• missing tool call chunks for openai logging (#657)
• arrow keys on windows (#661)

Other

• chore: bump version to 0.0.11 (#622)
• Add skipLoopDetection Configuration Option (#610)
• Chore/sync gemini cli v0.3.4 (#605)
• Enable tool call type coersion (#477)
• Vision model support for Qwen-OAuth (#525)

0.0.11 - 2025-09-12

Added

• Update the multilingual documentation links in the README (#536)
• Add Welcome Back Dialog, Project Summary, and Enhanced Quit Options (#553)
• Replace all Gemini CLI brand references with Qwen Code. (#588)

Changed

• cli: update OpenAI API key prompt with Bailian URL (#50)
• openaiContentGenerator (#501)

Fixed

• update OpenAIKeyPrompt test to expect Alibaba Cloud API URL (#560)
• resolve EditTool naming inconsistency causing agent confusion loops (#513)
• unexpected re-auth when auth-token is expired (#549)
• relax chunk validation to avoid unnecessary retry (#584)
• clear saved creds when switching authType (#587)
• tool calls ui issues (#590)

Other

• chore: add configurable cache control (#498)
• chore: pump version to 0.0.10 (#502)
• Terminal Bench Integration Test (#521)
• Fix E2E caused by Terminal Bench test (#529)
• Re-implement tokenLimits class to make it work correctly for Qwen and… (#542)
• Fix packages/cli/src/config/config.test.ts (#562)
• 🎯 Subagents Feature (#573)
• Make the ReadManyFiles tool share the "DEFAULT_MAX_LINES_TEXT_FILE" limit across files. (#563)
• Fix performance issues with SharedTokenManager causing 20-minute delays (#586)

0.0.10 - 2025-09-02

Documentation

• Add homebrew install (#474)

Other

• chore: bump version to 0.0.9 (#468)
• 🚀 Add Todo Write Tool for Task Management and Progress Tracking (#478)

• 🚀 Sync Gemini CLI v0.2.1 - Major Feature Update (#483)

0.0.9 - 2025-08-27

Added

• update /docs link (#438)

Fixed

• add explicit is_background param for shell tool (#445)
• sync token among multiple qwen sessions (#443)
• ambiguous literals (#461)

Other

• chore: pump version to 0.0.8 (#421)
• Sync upstream gemini-cli v0.1.21 (#398)
• Fix GitHub Workflows Configuration Issues (#451)
• Fix parallel tool use (#400)
• Fix race condition in submitQuery preventing tool response continuations (#458)
• use sub-command to switch between project and global memory ops (#450)
• 🔧 Miscellaneous Improvements and Refactoring (#466)

0.0.8 - 2025-08-22

Added

• use .geminiignore in grep tool (#349)
• Add deterministic cache control (#411)

Fixed

• revert trimEnd on LLM response content (#397)
• Critical Issues in v0.0.8-nightly.7 (#419)

Documentation

• Update security policy with Alibaba contact information (#390)

Other

• Chore/release 0.0.7 (#343)
• support: project/global save location option. (#368)
• doc: Add links to translated README versions (#171)
• Sync upstream gemini-cli v0.1.19 (#364)
• 🚀 Enhance Release Notes Generation with Previous Tag Detection (#394)
• Update Documentation Branding from Gemini CLI to Qwen Code (#391)
• Fix prompt re-submission (#392)
• Fix GitHub Workflows for Issue Triage (#396)
• Limit grep result (#407)

0.0.7 - 2025-08-15

Added

• sandbox: add GHA to build sandbox image (#262)
• prevent concurrent query submissions in useGeminiStream hook (#322)
• refactor web-fetch tool to remove google genai dependency (#340)

Fixed

• qwen logger exit handler setup (#325)
• seperate static QR code and dynamic spin components (#327)
• OpenAI tools (#328)
• custom API's trailing space and empty tool id issues (#326)

Other

• chore: add api request logger (#313)
• Sync with upstream gemini-cli v0.1.18 (#309)
• chore: bump version to 0.0.6 (#323)
• Migrate web search from Google/Gemini to Tavily API (#329)
• Update qwen-code-pr-review.yml (#342)

0.0.6 - 2025-08-12

Added

• add usage statistics logging for Qwen integration (#284)

Fixed

• rename make run-npx from gemini to qwen (#242)
• terminal flicker when waiting for login (#248)
• openaiContentGenerator (#283)
• 🐛 fix EPERM error when run qwen --sandbox in macOS (#293)

Other

• rename GEMINI.md to QWEN.md across the codebase (#235)
• Fix README.md: Replace /status command with /stats command in documen… (#266)
• Make /init respect configured context filename and align docs with QWEN.md (#274)
• chore: adjust workflow to run PR review (#297)
• Chore/pkg version (#298)

0.0.5 - 2025-08-08

Added

• Add systemPromptMappings Configuration Feature (#108)
• update /bug command to point to Qwen-Code repo (#154)
• add qwencoder as co-author (#207)
• oauth: add Qwen OAuth integration (#225)

Fixed

• resolve RadioButtonSelect array bounds crash and auth dialog navigation (#46)
• streaming token usage (#102)
• Enhanced OpenAI Usage Logging and Response Metadata Handling (#141)

Other

• pre-release: fix ci (#1)
• fix login preflight & sync with npm version (#55)
• add star history (#109)
• update: add info about modelscope-api (#116)
• Fix Default Model Configuration and Fallback Behavior (#142)
• Update: shrink/hard constrained token usage (#136)
• Fix E2E (#156)
• Fix Sandbox docker mode (#160)
• Support openrouter (#162)
• Update: add telemetry service (#161)
• Update README.md to clarify the requirement for using Modelscope inference API (#131)
• fix config (#163)
• fix release workflow (#172)
• sync gemini cli 0.1.15 (#175)
• fix e2e (#185)
• fix system md (#189)
• sync gemini cli 0.1.17 (#206)
• chore: remove google registry (#227)

0.0.4 - 2025-08-03

Other

• sync gemini cli 0.1.15 (#175)
• fix e2e (#185)
• fix system md (#189)

0.0.2 - 2025-08-01

See GitHub release for details.