Qwen Code cua-driver-rs-v0.20.6

Qwen CUA Driver prebuilt binaries (vendored under packages/cua-driver).

• macOS: codesigned + notarized universal binary + QwenCuaDriver.app
• Linux: unsigned (x86_64 + arm64, glibc 2.31 floor)
• Windows: unsigned UIAccess worker + native SDK payload (x86_64 + arm64); deployments must sign and trust the worker on the target machine before UIAccess activation
• Node.js: the same workflow publishes @qwen-code/cua-sdk after a clean install against these release assets and publishes the independently versioned @qwen-code/node-repl-mcp package after its MCP protocol smoke tests

Enable relative coordinates: CUA_DRIVER_RS_COORDINATE_SPACE=1 (default 0 = off; optional CUA_DRIVER_RS_COORDINATE_SCALE=1000).

Enable textual MCP payload filtering for affected model API routes: MCP_MODEL_PAYLOAD_FILTER=1 (default off).

<!-- Release notes generated using configuration in .github/release.yml at 6a211b465227dd35ad3c35a2dcb86a809590bcd0 -->

What's Changed

• fix(web-shell): realign two suites with shipped behavior to unbreak main CI (#11525) by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/11530

• feat(acp): register daemon-managed sessions in the session registry and let them send peer messages by @qqqys in https://github.com/QwenLM/qwen-code/pull/11488

• fix(ci): grant the packaging check the permissions its callee declares by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11532

• fix(dev): install hooks during worktree bootstrap by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11496

• test(core): cover bundled ConPTY fallback and terminal replies by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11555

• fix(review): dispatch review agents through parallel workflows by @wenshao in https://github.com/QwenLM/qwen-code/pull/11524

• feat(web-shell): render URLs in user messages as clickable links by @wenshao in https://github.com/QwenLM/qwen-code/pull/11464

• fix(vscode): preserve active file exclusion by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11568

• feat(serve): support 256 workspaces by default by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/11515

• fix(cli): drain notification queue when a turn errors out by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11559

• fix(review): keep host-trusted state out of the container's writable surface by @wenshao in https://github.com/QwenLM/qwen-code/pull/9983

• fix(acp): pass shell execution config to tool invocations by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11102

• refactor(dingtalk): remove obsolete background response aggregation by @qqqys in https://github.com/QwenLM/qwen-code/pull/11570

• feat(channels)!: remove message prefix filtering by @qqqys in https://github.com/QwenLM/qwen-code/pull/11571

• fix(ci): widen the replayed supersede-cede timeline margin to deflake Scripts by @wenshao in https://github.com/QwenLM/qwen-code/pull/11581

• test(review): pin the fold invariant the per-line escape rests on by @wenshao in https://github.com/QwenLM/qwen-code/pull/11533

• fix(autofix): charge regressions to the brake and gate test weakening by @wenshao in https://github.com/QwenLM/qwen-code/pull/10188

• fix(acp): Preserve submitted prompt provenance for auto recall by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/11455

• fix(serve): restore configured channels after restart by @qqqys in https://github.com/QwenLM/qwen-code/pull/11472

• feat(goal): start approved Web Shell proposals after their owning turn by @qqqys in https://github.com/QwenLM/qwen-code/pull/11360

• refactor(core): move the web_search DashScope client behind a backend interface by @qqqys in https://github.com/QwenLM/qwen-code/pull/11490

• fix(serve): load persisted MCP config after ACP preheat at startup by @kabishou11 in https://github.com/QwenLM/qwen-code/pull/11145

• fix(vscode-ide-companion): resolve diff paths once in DiffManager by @CanReader in https://github.com/QwenLM/qwen-code/pull/10494

• feat(sdk): add a peer endpoint so a program outside Qwen Code can join cross-session messaging by @qqqys in https://github.com/QwenLM/qwen-code/pull/11560

• feat(goal): stop a Goal at a turn or an active-time budget by @qqqys in https://github.com/QwenLM/qwen-code/pull/11457

• perf(core): drop the redundant history clone in session recovery planning by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/11595

• fix(core): free an exited web terminal's PTY resources at exit time by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/11572

• test(integration): assert successful /compress token reduction by @yu-xin-c in https://github.com/QwenLM/qwen-code/pull/11259

• fix(ci): retry transient E2E build artifact upload failures once (#11364) by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/11375

• feat(web-shell): route scheduled runs by model and group by @qqqys in https://github.com/QwenLM/qwen-code/pull/11396

• feat(web-shell): enrich browser notifications and open target sessions by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/11447

• feat(web-shell): improve composer and context usage presentation by @wenshao in https://github.com/QwenLM/qwen-code/pull/11509

• feat(live): add visual input, proactive assistance and memory by @wangxiongts in https://github.com/QwenLM/qwen-code/pull/11369

• fix(bridge): preserve child error detail on session explicit-close and kill paths (#11123) by @now-ing in https://github.com/QwenLM/qwen-code/pull/11190

• feat(review): focus automatic static navigation reviews by @wenshao in https://github.com/QwenLM/qwen-code/pull/11456

• fix(web-shell): persist cancelled responses and elapsed time by @ytahdn in https://github.com/QwenLM/qwen-code/pull/11608

• feat(subagents): add Codex executor and native agent builtins by @wenshao in https://github.com/QwenLM/qwen-code/pull/11474