Qwen Code Desktop v0.27.3 is a Alibaba release published 2026-11-13 with 170 listed changes. Summary and full notes below, extracted from the official notes.
One free email: every breaking change, retirement and price change across 23 coding agents. No spam, unsubscribe anytime.
Subscribe free← All Qwen Code releases and breaking changes
<!-- Release notes generated using configuration in .github/release.yml at ddeba2fd3436d86780dc90cf621e46e67542656e --> ## What's Changed * fix(ci): read prettier's whole ignore set at the base tip by @wenshao in https://github.com/QwenLM/qwen-code/pull/16729 * fix(cli,core): let an Agent Host retry an enrollment whose answer was lost by @wenshao in https://github.com/QwenLM/qwen-code/pull/16722 * fix(channels): shape-check authorExclude on the inbound task record by @wenshao in https://github.com/QwenLM/qwen-code/pull/16740 * feat(managed-agent): import Skills from the original Runtime by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16638 * docs: clarify tool bridge requirements when migrating from tools.core by @521Peter in https://github.com/QwenLM/qwen-code/pull/16750 * fix(ci): annotate a truncated tree and pin the encoded tree ref by @wenshao in https://github.com/QwenLM/qwen-code/pull/16747 * refactor(managed-agent): give JDBC stores names distinct from their interfaces by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16759 * fix(ci): stop the Ctrl+C and Java heap failures that block fork PRs by @TianYuan1024 in https://github.com/QwenLM/qwen-code/pull/16150 * fix(core): decode ANSI-C words and the -c option terminator with bash's rules by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16540 * feat(cli): thread /btw session history and decouple its request lifecycle by @sorrycc in https://github.com/QwenLM/qwen-code/pull/16408 * refactor(managed-agent): share Hosted Harness feature tokens between connectors by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16763 * fix(ci): stop stale-base updates from restarting timed-out PR reviews by @wenshao in https://github.com/QwenLM/qwen-code/pull/16749 * fix(web-shell): bump @datafe-open/markdown-chart to 0.4.1 by @zhangxy-zju in https://github.com/QwenLM/qwen-code/pull/16021 * ci: reduce all sharded CI lanes to two shards by @wenshao in https://github.com/QwenLM/qwen-code/pull/16764 * feat(managed-agent): add Tool v4 execution measurements by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16498 * test(core): keep the background renewal out of the byte-offset title case by @wenshao in https://github.com/QwenLM/qwen-code/pull/16756 * perf(memory): drop the bundled auto-memory reference and keep the resident policy by @qqqys in https://github.com/QwenLM/qwen-code/pull/16739 * perf(github): poll with conditional requests so unchanged data costs no quota by @wenshao in https://github.com/QwenLM/qwen-code/pull/16741 * feat(serve): add PR-watch host discovery and ordered shutdown by @wenshao in https://github.com/QwenLM/qwen-code/pull/16732 * fix(web-templates): lift the export renderer budget to the merged size by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16774 * feat(cli): send OpenTUI's model rounds through the shared agent loop by @qqqys in https://github.com/QwenLM/qwen-code/pull/16743 * feat(mods): run model completions from Mods by @qqqys in https://github.com/QwenLM/qwen-code/pull/16737 * feat(managed-agent): add private BTW and session recap by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16758 * fix(cli): let idle-only slash commands run from an idle session by @qqqys in https://github.com/QwenLM/qwen-code/pull/16731 * feat(core): Goal verifier verdict inherits the session output language by @4ekuct25 in https://github.com/QwenLM/qwen-code/pull/15806 * fix(web-shell): prioritize slash menu names by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16790 * refactor(cli): single-source the Managed Runtime worker route surface by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16788 * docs(managed-agent): design Managed /loop support (schedules and wakeups) by @wenshao in https://github.com/QwenLM/qwen-code/pull/16792 * fix(cli): notify when a background agent or workflow parks an approval by @qqqys in https://github.com/QwenLM/qwen-code/pull/16800 * feat(mods): let Mods rewrite reply rows in the Ink terminal by @qqqys in https://github.com/QwenLM/qwen-code/pull/16794 * feat(ci): raise the self-hosted autofix agent budget to 240 minutes by @wenshao in https://github.com/QwenLM/qwen-code/pull/16777 * fix(cli): defer unlisted subscription model errors to send by @liuxuely0626 in https://github.com/QwenLM/qwen-code/pull/16798 * feat(web-shell): configure the Email channel from the channel manager by @wenshao in https://github.com/QwenLM/qwen-code/pull/16746 * feat(cli): run OpenTUI's tool batches on the shared agent loop by @qqqys in https://github.com/QwenLM/qwen-code/pull/16789 * fix(core): write auto session titles in the conversation's language by @qqqys in https://github.com/QwenLM/qwen-code/pull/16806 * feat(mods): keep each plugin's values between sessions with $.store by @qqqys in https://github.com/QwenLM/qwen-code/pull/16791 * fix(cli): show a foreground agent's pending approval in the live panel by @qqqys in https://github.com/QwenLM/qwen-code/pull/16809 * fix(cli): only enable sandbox debugging for explicit DEBUG values by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16805 * fix(deps)!: upgrade dependencies shipped by SDK and Web Shell by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16452 * docs(managed-agent): Managed scheduling designs — loop review fixes and external schedulers by @wenshao in https://github.com/QwenLM/qwen-code/pull/16796 * fix(core): attribute gh pr create in the directory the command runs in by @wenshao in https://github.com/QwenLM/qwen-code/pull/16725 * chore(managed-agent): remove unused constructors and helpers by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16822 * fix(core): keep reacquireWriter from acquiring behind a seal by @wenshao in https://github.com/QwenLM/qwen-code/pull/16726 * perf(managed-agent): cheapen proof copies, CSI lookup and startup polling by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16828 * refactor(serve): extract the session restore core for in-process callers by @wenshao in https://github.com/QwenLM/qwen-code/pull/16810 * feat(web-shell): let managed sessions reuse the host shell by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16324 * feat(managed-agent): keep the channel delivery ledger segment-faithful by @wenshao in https://github.com/QwenLM/qwen-code/pull/16797 * feat(channels): gate review/verify bound delivery on active watches by @wenshao in https://github.com/QwenLM/qwen-code/pull/16772 * feat(web-shell): configure the WeChat channel with in-browser QR login by @wenshao in https://github.com/QwenLM/qwen-code/pull/16736 * fix(core): reject recycled PIDs in runtime session status by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16128 * docs(design): correct the wakeDormant delivery claim to a latency gap by @wenshao in https://github.com/QwenLM/qwen-code/pull/16834 * perf(managed-agent): remove redundant read-path queries by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16827 * feat(channels): add a QR-login management contract and the QQ config surface by @wenshao in https://github.com/QwenLM/qwen-code/pull/16724 * feat(managed-agent): expose channel delivery segments in the public read model by @wenshao in https://github.com/QwenLM/qwen-code/pull/16807 * feat(web-shell): show session speed and cache hit under the composer by @qqqys in https://github.com/QwenLM/qwen-code/pull/16815 * fix(channels): let a config field declare an inclusive lower bound by @wenshao in https://github.com/QwenLM/qwen-code/pull/16783 * fix(managed-agent): release the MCP owner when a Session closes by @wenshao in https://github.com/QwenLM/qwen-code/pull/16728 * fix(managed-agent): harden Harness error-code parsing, host reader and exception causes by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16826 * docs(providers): document Wiro AI under OpenAI-compatible providers by @EminBudak in https://github.com/QwenLM/qwen-code/pull/16850 * docs(serve): describe workspace Skill settings writes by name by @521Peter in https://github.com/QwenLM/qwen-code/pull/16833 * refactor(managed-agent): keep asynchronous admission as the only publication mode by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16840 * feat(web-shell): show unfinished background tasks as chips above the composer by @qqqys in https://github.com/QwenLM/qwen-code/pull/16813 * feat(channels): localize bound-session event prompts by UI language by @wenshao in https://github.com/QwenLM/qwen-code/pull/16773 * test(cli): cover OpenTUI's steering input and retry on the shared agent loop by @qqqys in https://github.com/QwenLM/qwen-code/pull/16820 * feat(managed-mcp): propose live-only selection capacity by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/15161 * fix(web-shell): compact approval picker and flag full access by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16874 * fix(ci): preserve live ephemeral runners past their age threshold by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16871 * test(managed-agent): make the channel recording Harness faithful to the funnel by @wenshao in https://github.com/QwenLM/qwen-code/pull/16843 * fix(web-shell): give the WeChat QR mint the budget the daemon allows it by @wenshao in https://github.com/QwenLM/qwen-code/pull/16838 * docs: add Qwen Code 1.0 Preview entry to README by @0xvincent in https://github.com/QwenLM/qwen-code/pull/16872 * fix(cli): tolerate missing container $HOME and create sandbox home dir by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16878 * fix(web-shell): reconcile acknowledged managed model changes by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16869 * fix(ci): settle the parity approval wait and restore checkstyle import order by @wenshao in https://github.com/QwenLM/qwen-code/pull/16881 * docs: format README with the repository Prettier version by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16887 * test(cli): deflake the hosted harness message-stop test (#16853) by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/16855 * fix(ci): install fonts in the ephemeral runner image (#16863) by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/16864 * docs(channels): name the WeChat conversation the bot actually reads by @wenshao in https://github.com/QwenLM/qwen-code/pull/16888 * feat(managed-agent): Add external extension source and archive storage SPIs by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16882 * test(web-shell): disambiguate the sidebar Collapse button from Collapse all workspaces by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16905 * ci(desktop): re-pin the RHEL 10 build root to a current CentOS Stream 10 digest by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16897 * test(web-shell): expect the subscription activation's model-persist opt-out meta by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16904 * test(web-shell): start the composer context mobile hover check from neutral focus by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16907 * fix(ci): drain ephemeral runners before shutdown by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16892 * refactor(managed-agent): scope CSI retirement evidence to its own session by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16910 * feat(mods): let Mods rewrite reply rows in the OpenTUI terminal by @qqqys in https://github.com/QwenLM/qwen-code/pull/16873 * fix(web-shell): keep explanation popover at its intended width by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16916 * fix(managed-agent): read the frozen CSI commit body with its own size bound by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16923 * test(generated-ui): add the #14019 acceptance evaluation by @qqqys in https://github.com/QwenLM/qwen-code/pull/16811 * feat(serve): record the slim listing's states and carry its keys to the watch by @wenshao in https://github.com/QwenLM/qwen-code/pull/16847 * fix(daemon): allow manual forget when auto memory is disabled by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16913 * fix(web-shell): align approval explanation and actions by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16914 * docs(managed-agent): sync the reply-loss gate design to terminal-drain release by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16898 * fix(web-shell): group session sorting with multi-select by @deggs7 in https://github.com/QwenLM/qwen-code/pull/16776 * feat(channels): add DingTalk webhook delivery with configured targets by @jkxiao in https://github.com/QwenLM/qwen-code/pull/16802 * test(web-shell): fix the nightly Browser Regression cluster by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16919 * fix(web-shell): raise build heap to 4096MB to fix companion release macOS OOM by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16902 * refactor(managed-agent): complete broker facade decomposition by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16922 * refactor(managed-agent): remove uncalled CSI admission and workspace catalog code by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16928 * feat(web-shell): expand sidebar session lists ten at a time by @wenshao in https://github.com/QwenLM/qwen-code/pull/16830 * docs(channels): say what the WeChat pairing reply actually is by @wenshao in https://github.com/QwenLM/qwen-code/pull/16930 * feat(managed-agent): OSS externalization of oversized session resources and journal compaction by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16868 * fix(channels): retry stored GitHub replies while the channel runs by @wenshao in https://github.com/QwenLM/qwen-code/pull/16839 * fix(cli): count daemon-persisted routes in channel status by @wenshao in https://github.com/QwenLM/qwen-code/pull/16929 * fix(managed-agent): honour the declared base64 bounds and drop an orphaned import by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16934 * feat(core): add event-bus primitives for the agent loop by @wenshao in https://github.com/QwenLM/qwen-code/pull/16935 * fix(managed-agent): align model selection with DAC by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16890 * fix(ci): heal pool runner leftovers that break checkout and corepack by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16939 * fix(core): honour permissions.ask rules in auto-edit mode by @qqqys in https://github.com/QwenLM/qwen-code/pull/16932 * fix(web-shell): admit GitLab as a repository-only channel by @wenshao in https://github.com/QwenLM/qwen-code/pull/16845 * fix(managed-agent): map the resend identity conflict to 409 and pin the projection wire shape by @wenshao in https://github.com/QwenLM/qwen-code/pull/16846 * refactor(managed-agent): extract binding admission and fence engine by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16936 * test(ci): pin the widened lint-lane gates that had no assertion by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16950 * ci: bake the pool's browsers into its image and cap the container typecheck heap by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16951 * fix(managed-agent): give the CSI retirement size bounds a single owner by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16956 * test(integration): deflake the TUI loop parity startup date by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/16961 * fix(ci): re-baseline ci.yml workflow size after merged CI growth by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/16978 * refactor(managed-agent): remove unused CSI checkpoint snapshot and log readers by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16957 * chore(managed-agent): finish the CSI bound conversion and name it once by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16981 * fix(managed-agent): fence concurrent journal pack publication with a builder lease by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16983 * ci: route web shell PR jobs to ephemeral runners by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16917 * feat(web): configure DingTalk webhooks and scheduled delivery targets by @jkxiao in https://github.com/QwenLM/qwen-code/pull/16804 * fix(web-shell): center the session timeline vertically by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16912 * refactor(managed-agent): retire the InMemory runtime broker repositories by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16953 * feat(web-shell): support host model picker content by @callmeYe in https://github.com/QwenLM/qwen-code/pull/16933 * ci: re-run unit suites the worker RPC budget failed on a contended host by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16945 * feat(core): re-express the agent loop on the event bus (opt-in) by @wenshao in https://github.com/QwenLM/qwen-code/pull/16938 * fix(web-shell): cancel the QQ login session when the editor closes by @wenshao in https://github.com/QwenLM/qwen-code/pull/16841 * fix(core): publish file-history backups atomically by @wenshao in https://github.com/QwenLM/qwen-code/pull/16926 * fix(cli): never terminally release a config-scoped Hook owner on a detach by @wenshao in https://github.com/QwenLM/qwen-code/pull/16594 * perf(managed-agent): gate the reconcile record read and batch the segment plan by @wenshao in https://github.com/QwenLM/qwen-code/pull/16865 * feat(verify-pr): deepen the verify checkout to the PR's history by @wenshao in https://github.com/QwenLM/qwen-code/pull/16781 * docs(managed-agent): drop removed WorkspaceCsiCheckpointSnapshotStore from the deletion precondition by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/16984 * feat(cli): enable the event-bus agent loop on headless behind an experimental setting by @wenshao in https://github.com/QwenLM/qwen-code/pull/16994 * docs(managed-agent): record why the gate cannot be a required check yet by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16955 * ci: refuse an unrunnable host before the unit suites start by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16987 * test(core): pin the default executor's scheduler and runtimeView passthrough by @wenshao in https://github.com/QwenLM/qwen-code/pull/17009 * refactor(managed-agent): settle Session rename synchronously under one display authority by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/17006 * docs(design): cross-link the event-bus loop and refresh the unified-loop status by @wenshao in https://github.com/QwenLM/qwen-code/pull/17014 * docs(agents): name the pnpm install in the Browser Use dev note by @qqqys in https://github.com/QwenLM/qwen-code/pull/17019 * feat(core): send AgentCore's model rounds through the shared agent loop by @qqqys in https://github.com/QwenLM/qwen-code/pull/16880 * docs(agents): import AGENTS.md from CLAUDE.md by @qqqys in https://github.com/QwenLM/qwen-code/pull/17017 * ci: narrow lint gate freshness to the lint slice, and close its follow-ups by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/16946 * fix(core): resolve system prompts for the request model by @callmeYe in https://github.com/QwenLM/qwen-code/pull/17023 * test(cli): pin the host machine-id in SSH bootstrap process tests (#16975) by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/16976 * test(ci): read the lint freshness gate's fatality through the YAML parser by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/17028 * fix(cli): refuse an unreadable host identity instead of leaking ENOENT by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/17034 * fix(web-shell): 仅首页时隐藏多余导航栏并保留底部操作 by @chiga0 in https://github.com/QwenLM/qwen-code/pull/17027 * fix(browser-use): move the bundled zod to 4.6.6 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/17022 * fix(web-shell): reflect welcome model selection in settings by @deggs7 in https://github.com/QwenLM/qwen-code/pull/16819 * fix(permissions