OpenClaw v2026.6.8 - 更新日志

## 2026.6.8 ### 亮点 - **更丰富的渠道投递：** Telegram 和 WhatsApp 更稳定：Telegram 可渲染带表格、列表、可展开引用块、保留有意换行及 CLI 回复的结构化文本，而 WhatsApp 现在能遵循配置的 ACP 绑定。(#92679, #93164, #84082, #89421, #92513) 感谢 @obviyus, @vincentkoc, @jzakirov, @spacegeologist, @TurboTheTurtle 和 @mcaxtr。 - **更可靠的代理运行：** 账户范围的 DM 发送、生成的媒体补全、自动回复消息工具的最终回复、重置存档回退读取、重启中止、暂停的子代理、会话身份提示均保持在正确的恢复路径上。(#92788, #91246, #92879, #91357, #92631, #92468) 感谢 @yetval, @TurboTheTurtle, @masatohoshino, @CadanHu, @vincentkoc, @ooiuuii, @openperf, @zhangguiping-xydt 和 @QQSHI13。 - **更安全的模型路由：** 新增 GLM-5.2 和 Claude Haiku 4.5 目录支持，使用标准化提供商 ID、托管 SecretRef 鉴权、有限的模型浏览以及更安全的 OpenAI/Anthropic 工具架构恢复。(#92796, #90116, #92627, #90686, #92247, #92941) 感谢 @arkyu2077, @liuhao1024, @lijenhsin, @rohitjavvadi 和 @samson910022。 - **实用的使用量页脚：** `/usage` 和回复载荷钩子现在拥有原生的完整页脚渲染器、默认模板、固定小数格式、凭据感知的限制、更好的部分计数处理，以及针对损坏模板的警告（而非静默输出错误）。(#92657, #89835, #89629) 感谢 @Marvinthebored。 - **可预测的网页搜索默认值：** 无密钥提供商（如 Parallel Free、DuckDuckGo、Ollama、Codex Hosted Search）保持为显式选择加入，而非自动回退，避免意外。(#93616) 感谢 @davemorin 和 @vincentkoc。 - **更平静的 UI 和移动会话：** 工作区文件默认折叠、WebChat 回滚在流式传输中幸存、桌面会话选择器保持可交互、重置参数在调度中幸存、iOS 重新连接过时的前台网关。(#92779, #92622, #92705, #91353, #92552) 感谢 @shakkernerd, @TurboTheTurtle, @NianJiuZst, @zhouhe-xydt 和 @Solvely-Colin。 - **弹性的记忆与状态：** 过大的 OpenAI 嵌入批次在 431 秒前拆分、QMD 搜索在瞬态模式下保持可用、SQLite 在 NFS 卷上避免 WAL、完整重建保留回滚/缓存恢复。(#92650, #92618, #92639, #91247, #92881) 感谢 @mushuiyu886, @BrettHamlin, @zhbcher, @TurboTheTurtle, @Takhoffman, @849261680 和 @TSHOGX。 ### 变更 - 提供商/模型：添加 GLM-5.2 支持和 Claude Haiku 4.5 目录条目，同时保持提供商限定的模型 ID 在 OpenRouter 和 Google Vertex 路径上标准化。(#92796, #90116, #92627, #91218) 感谢 @arkyu2077, @liuhao1024 和 @bymle。 - 网页搜索：保持无密钥提供商（如 Parallel Free、DuckDuckGo、Ollama、Codex Hosted Search）为显式选择加入，而非在未配置 API 支持的提供商时自动选择它们。(#93616) 感谢 @davemorin 和 @vincentkoc。 - 渠道插件：推出 Telegram 富消息投递和 WhatsApp ACP 绑定支持，包括保留有意换行、富提示交接给 CLI 后端，以及用于更丰富草稿的传输装置。(#92679, #93164, #92513) 感谢 @obviyus 和 @TurboTheTurtle。 - 代理命令：在 CLI 后端会话中支持 `/btw`，并将 CLI 使用错误退出归类为使用失败而非成功运行。(#92669, #92162) 感谢 @joshavant 和 @Pandah97。 - 使用量钩子：添加内置完整页脚渲染、默认页脚模板、每次对话的使用状态、凭据感知的限制，以及使用量条模板的固定小数格式化。(#92657, #89835, #89629) 感谢 @Marvinthebored。 ### 修复 - 渠道与投递：保留账户范围 DM 渠道发送策略、Telegram 中富消息的有意换行和状态输出、富文本 Telegram 最终回复、富文本 Telegram 表格和列表、Telegram 线程创建 CLI 重映射、Feishu 在持久化绑定重用后的动态代理路由、Slack 出站 `message_sent` 钩子、贡献的消息工具架构可选性、同渠道生成的媒体补全，以及围绕代理对和 Infinity 限制的渠道分块。(#92788, #93164, #92679, #89421, #89943, #42837, #92814, #91137, #91246, #92735) 感谢 @yetval, @obviyus, @spacegeologist, @rishitamrakar, @liuhao1024, @lundog, @TurboTheTurtle 和 @yhterrance。 - Gemini CLI：在隔离的 Gemini CLI 运行时主目录中使用选定的 OpenClaw OAuth/API 密钥鉴权配置文件，防止环境中的 Google 机器凭据覆盖所选配置文件。(#88748) 感谢 @jason-allen-oneal 和 @shakkernerd。 - Discord：为生成的自动线程标题设置 60 秒超时和 4,096 个令牌的推理模型输出预算，并限制在所选模型的输出上限内。(#64734) 感谢 @hanamizuki。 - 代理、cron 和网关运行时：在重启中止前标记活动主会话、暂停其终端也发送中止信号的子代理运行、通过提供商/模型回退限制受信子代理思维覆盖、保留产生的媒体补全、通过自动回复传递渠道消息工具最终回复同时隐藏内部投递提示、在活动异步转录缺失时恢复重置存档回退读取、去重主会话心跳事件、在运行时提示中暴露会话身份、拒绝未知的 OpenAI 代理选择器、保留生成的媒体补全、斜杠命令块回复和 WebChat 中的轨迹导出命令，并需要管理员权限才能使用 HTTP 会话/模型覆盖界面。(#91357, #92631, #92412, #92146, #92879, #91287, #92468, #92510, #91246, #92651, #92646) 感谢 @ooiuuii, @openperf, @IWhatsskill, @masatohoshino, @CadanHu, @ZengWen-DT, @zhangguiping-xydt, @TurboTheTurtle 和 @oiGaDio。 - 提供商与模型重放：保留无存储 OpenAI 响应的重放兼容性、恢复无效的 OpenAI 推理签名和泛化的 Anthropic 思维签名重放错误、将 OAuth 图像默认通过 Codex 路由到符合条件的 OpenAI 配置文件、在 Copilot 中禁用 Claude 4.5 的急切工具流式传输、隔离不可读和后钩子 OpenAI/Anthropic 系列工具架构而不扩大允许的工具选择、向 LM Studio 二进制思维模型发送显式关闭思维请求、为 SecretRef 模型条目遵守配置文件鉴权、限制模型浏览、在运行时需要裸 ID 时去除提供商前缀，并暴露嵌套嵌入提取失败。(#90706, #92941, #92201, #92916, #92824, #75393, #92908, #92921, #92928, #92002, #90686, #92247, #92627, #91218, #92628) 感谢 @snowzlm, @mmyzwl, @CarlCapital, @bek91, @Kailigithub, @vincentkoc, @rohitjavvadi, @samson910022, @nxmxbbd, @liuhao1024, @bymle 和 @mushuiyu886。 - 记忆、状态、诊断与配置：拆分头部过大的嵌入批次、在瞬态模式下保持 QMD 记忆搜索启用、在 NFS 卷上避免 SQLite WAL、在卡会话警告退避之外保留恢复调度、保留完整重建回滚/缓存恢复，并将原始 Memory Wiki 源页面视为源证据。(#92650, #92618, #92639, #91247, #92752, #92881, #59137, #92876) 感谢 @mushuiyu886, @TurboTheTurtle, @849261680, @gnanam1990, @TSHOGX 和 @arlen8411。 - UI/移动端/TUI：保留仪表盘会话父级谱系、WebChat 回滚、重置软命令参数、侧边栏会话选择器可交互性、工作区文件默认折叠、解析的 `/model` 确认引用、过时的前台 iOS 网关重新连接，以及在继承 stdin 子进程退出后暂停父级 stdin。(#90658, #92622, #91353, #92705, #92779, #92773, #92552, #93159) 感谢 @luoyanglang, @TurboTheTurtle, @zhouhe-xydt, @NianJiuZst, @shakkernerd, @NarahariRaghava, @Solvely-Colin 和 @fuller-stack-dev。 - 插件与更新：修复托管插件安装和更新中缺失的必要平台包，包括遗漏的 Codex 平台二进制文件。 - 依赖项：将 Hono 更新至 4.12.25，使已发布的 OpenClaw 和 ACPX 包使用修补后的运行时。 - 发布可靠性：延长慢速网关/完整套件看门狗、在受限时拆分本地完整套件分片、将 Telegram RTT 采样纳入实时 QA 证据、简化 QA 评分卡映射围绕规范覆盖 ID、在固定预期 SHA 的同时保留发布检查子引用、扩大实时 OpenAI TTS 预算以应对较慢的提供商响应，并避免未解决的最新标签更新的虚假降级提示。(#92550, #92558, #92911) 感谢 @RomneyDa 和 @Andy312432。 ### 完整贡献记录此审核记录涵盖了 v2026.6.6..v2026.6.8 的完整历史：168 个合并的 PR。生成清单也提供了直接提交作为编辑输入；上述分组说明优先考虑用户影响。 #### 拉取请求 - **PR #92144** 修复(cron)：在 cron.status 中报告 SQLite 存储路径而不是旧的 jobs.json。关联 #91766。感谢 @liuhao1024 和 @AaronFaby。 - **PR #92175** 修复(channel)：加强本地设置信任。感谢 @hxy91819。 - **PR #92073** 修复：处理显式的静默助手回复。感谢 @sallyom。 - **PR #92087** 修复(docker)：在镜像中包含 QA Lab 运行时。感谢 @jesse-merhi。 - **PR #92387** 修复(anthropic-vertex)：停止在传输预算的载荷上重新标记 cache_control。感谢 @openperf 和 @Takhoffman。 - **PR #92229** 修复 doctor 预览渠道 SecretRef 解析。感谢 @joshavant。 - **PR #92225** 修复禁用心跳的一次性 cron 重试。感谢 @joshavant。 - **PR #92265** 修复配置的 DeepSeek 模型传输继承。感谢 @joshavant。 - **PR #92226** CLI 支持的 /btw 回退失败时关闭。感谢 @joshavant。 - **PR #92231** 修复抑制的心跳承诺投递。感谢 @joshavant。 - **PR #92280** 修复(agents)：将结构化不支持的模型错误分类。感谢 @joshavant。 - **PR #92276** 修复 OTLP 日志跟踪关联。感谢 @joshavant。 - **PR #92282** 修复(update)：移交 Linux 服务自动更新。感谢 @joshavant。 - **PR #92235** 修复：解析托管 SecretRef 提供商鉴权。感谢 @joshavant。 - **PR #92293** 修复提供商静态模型回退解析。感谢 @joshavant。 - **PR #92343** 修复(agent)：在源消息工具回复后继续。感谢 @joshavant。 - **PR #92350** 修复(codex)：保留记忆提示注册。感谢 @rubencu。 - **PR #92290** 修复：澄清网关 SecretRef 鉴权诊断。感谢 @joshavant。 - **PR #92286** 修复：修复被拒绝的 Anthropic 思维重放。感谢 @joshavant。 - **PR #92281** 修复 Telegram 假脱机缓冲重放。感谢 @joshavant。 - **PR #92416** 修复(outbound)：遵循顶级图像参数作为发送媒体源 (#92407)。感谢 @xydigit-sj 和 @ichirokyoto。 - **PR #92508** 修复(sandbox)：从物化路径渲染 CLI 技能提示。感谢 @brokemac79。 - **PR #92540** chore：修复 esbuild 生产审计失败。感谢 @RomneyDa。 - **PR #91484** 添加 QA 证据工件输出。感谢 @RomneyDa。 - **PR #91500** 添加 QA 评分卡分类验证。感谢 @RomneyDa。 - **PR #92554** feat(moonshot)：添加 Kimi K2.7 Code 支持。 - **PR #92396** 修复(moonshot)：在助手工具调用重放消息上回填 reasoning_content。关联 #71491。感谢 @xialonglee 和 @RoseKongPS。 - **PR #92566** 修复生命周期超时清理在领导退出后。感谢 @RomneyDa。 - **PR #88993** 暴露分页渠道动作结果。感谢 @fuller-stack-dev。 - **PR #90326** 修复(fireworks)：通过核心从 plugin.json 解析目录模型参数。感谢 @obuchowski。 - **PR #86629** 修复(doctor)：对不受信任的外部 Discord 插件发出警告。关联 #83212。感谢 @brokemac79 和 @cdeyoung67。 - **PR #90242** 修复(providers)：跳过不可读的 Mistral 工具架构。感谢 @vincentkoc。 - **PR #92498** 修复(reply)：镜像同渠道 Slack 最终回复。关联 #92489。感谢 @TurboTheTurtle 和 @TalkingHeadsJed。 - **PR #92083** 修复(channels)：为被吞没的进度草稿启动错误设置默认边界记录器。感谢 @hansraj316。 - **PR #92031** 修复(channels)：使定时器触发的进度草稿启动错误可观察。感谢 @hansraj316。 - **PR #92564** 修复(agents)：隔离无效插件模型目录 [AI 辅助]。关联 #92553。感谢 @tangtaizong666 和 @fxstein。 - **PR #89827** docs：UX-013 — 设计系统文档。感谢 @BunsDev。 - **PR #89615** feat(ui)：隐藏空的工作板列。感谢 @BunsDev。 - **PR #89822** 修复(a11y)：B-1+B-2+B-3 — 对比度、焦点状态、最小字号。感谢 @BunsDev。 - **PR #92618** 修复 #92218：memory_search 工具在 QMD 后端被禁用。感谢 @mushuiyu886 和 @zhbcher。 - **PR #92608** docs(gateway)：在健康检查文档中添加运行时间监控指南（修复 #55768）。感谢 @liuhao1024 和 @faahim。 - **PR #92605** 修复(docs)：将 Windows Hub 下载链接固定到 v2026.6.5。关联 #92470。感谢 @lzyyzznl 和 @arjkul。 - **PR #92593** #92589：修复(internal-runtime-context)：将提示前言运行时上下文主体包装在分隔符中。感谢 @zhangqueping 和 @jovi2014-cyber。 - **PR #92606** 从 QA 套件运行 Vitest 和 Playwright 场景。感谢 @RomneyDa。 - **PR #89629** feat(hooks)：reply_payload_sending 上的每次对话 usageState。感谢 @Marvinthebored。 - **PR #89835** feat(usage)：原生模板化 /usage 完整页脚渲染器。感谢 @Marvinthebored。 - **PR #92247** 修复(models)：限制 /models 和模型列表目录加载。关联 #91809。感谢 @samson910022 和 @syfvb。 - **PR #90686** 修复(gateway)：为 SecretRef 模型条目遵循配置文件鉴权。关联 #90685。感谢 @rohitjavvadi。 - **PR #92651** 修复：需要管理员权限才能进行 HTTP 会话终止。感谢 @steipete-oai。 - **PR #92652** test(models)：稳定插件鉴权标记装置。 - **PR #89438** 修复(slack)：当渠道映射按名称而非渠道 ID 键控时发出警告。关联 #81665。感谢 @Alix-007 和 @cjalden。 - **PR #92631** 修复(agents)：暂停终端也发送中止信号的子代理运行。关联 #92448。感谢 @openperf 和 @aleps001。 - **PR #92622** 修复(ui)：在流式传输期间保留 WebChat 回滚。感谢 @TurboTheTurtle。 - **PR #92627** 修复(openrouter)：在 normalizeResolvedModel 钩子中去除模型 ID 上的 openrouter/ 前缀（修复 #92611）。感谢 @liuhao1024 和 @lijenhsin。 - **PR #92146** 修复(cron)：保留产生的媒体补全。感谢 @IWhatsskill。 - **PR #90116** 修复：添加 Claude Haiku 4.5 静态目录条目。关联 #90088。感谢 @arkyu2077 和 @maaron34。 - **PR #91137** 修复(channels)：保持贡献的消息工具架构属性为可选。感谢 @lundog。 - **PR #75393** 修复(copilot)：禁用 Claude 4.5 的急切工具流式传输。关联 #75348。感谢 @Kailigithub 和 @finchinslc。 - **PR #92628** 修复 #73713：暴露嵌套嵌入提取失败。感谢 @mushuiyu886 和 @crsnpalmer-art。 - **PR #89943** 修复(slack)：在出站投递时发送 message_sent 钩子（镜像 Telegram）。关联 #89942。感谢 @rishitamrakar。 - **PR #92668** 修复(docs)：在跳过前完成 i18n 后处理。感谢 @hxy91819。 - **PR #92673** 修复：分离图像设置和请求超时语义。感谢 @hxy91819。 - **PR #92639** 修复(memory)：在瞬态 qmd 模式下保持 memory_search。感谢 @TurboTheTurtle 和 @Takhoffman。 - **PR #92705** 修复(ui)：在桌面工作台上方恢复侧边栏会话选择器交互性。感谢 @NianJiuZst。 - **PR #92669** feat：在 CLI 后端会话中支持 /btw。感谢 @joshavant。 - **PR #91357** 修复(gateway)：在重启中止前标记活动主会话。感谢 @ooiuuii。 - **PR #92552** 修复(ios)：强制过时的前台网关重新连接。感谢 @Solvely-Colin。 - **PR #92752** 修复(diagnostics)：将恢复调度排除在卡会话警告退避之外。感谢 @gnanam1990 和 @Takhoffman。 - **PR #92766** 澄清 before_install 钩子范围。感谢 @sallyom。 - **PR #92513** 遵循 WhatsApp 配置的 ACP 绑定。感谢 @TurboTheTurtle 和 @mcaxtr。 - **PR #92796** feat(providers)：添加 GLM-5.2 支持。 - **PR #92807** 修复(heartbeat)：将出站镜像路由到隔离的会话密钥。感谢 @agent-merkava。 - **PR #92745** 修复(memory)：解释跳过的短期回忆命中。感谢 @mushuiyu886。 - **PR #92488** 修复(gateway)：在 /v1/responses 上转发仅图像输入（与聊天补全一致）。感谢 @s554097550。 - **PR #92604** 修复(status)：避免上下文百分比的累积使用量。感谢 @ashishpatel26。 - **PR #92547** 修复(nodes)：暴露待重新批准诊断。感谢 @fuller-stack-dev。 - **PR #92690** 修复(doctor)：当存储已迁移时避免误报的遗留 cron 存储警告（修复 #92683）。感谢 @liuhao1024 和 @motteman。 - **PR #92806** 修复(telegram)：当用户显式配置非 ipv4first dnsResultOrder 时跳过 IPv4 回退（修复 #41671）。感谢 @liuhao1024 和 @vincentkoc 和 @leandroirani933-ctrl。 - **PR #92778** 修复(macos)：延迟 isOverflowing 变异以打破 SwiftUI 渲染循环（修复 #43480）。感谢 @liuhao1024 和 @vincentkoc 和 @gdiab。 - **PR #92795** 修复(gateway)：使用 resolveNonNegativeNumber 处理 totalTokens，显示 0 而不是 ?（修复 #43009）。感谢 @liuhao1024 和 @vincentkoc 和 @ltxy12138-ai。 - **PR #92746** 修复(gateway)：在插件完成期间保留活动运行。感谢 @scotthuang 和 @vincentkoc。 - **PR #92820** UI：本地化日志选项卡标签。感谢 @rubensfox20。 - **PR #61080** UI：本地化日志硬编码标签。感谢 @rubensfox20。 - **PR #92825** 修复(telegram)：在给通用回调数据加前缀时保留命令回调。关联 #54909。感谢 @hnshah 和 @timt80。 - **PR #54962** 修复(telegram)：为内联按钮回调添加 'callback_data:' 前缀。关联 #54909。感谢 @hnshah 和 @timt80。 - **PR #87060** 修复(copilot)：从重放中丢弃思考块。关联 #81520。感谢 @giodl73-repo 和 @warcold。 - **PR #81534** 修复(github-copilot)：从最新的助手轮次中剥离思考块 (#81520)。感谢 @SymbolStar 和 @warcold。 - **PR #92834** feat(browser)：将 --labels 覆盖扩展到全页和元素捕获。感谢 @hxy91819。 - **PR #92095** 修复 #92039：[Bug]：WhatsApp 登录在鉴权持久化之前报告成功，因此 Docker 重建/升级可能强制重新链接。感谢 @zhangguiping-xydt 和 @dinorastoder。 - **PR #92801** 修复(stale)：从陈旧生命周期中豁免 ClawSweeper 可操作标签（修复 #89564）。关联 #78640, #81078, #81122。感谢 @liuhao1024 和 @brokemac79 和 @MilleniumGenAI 和 @Jackten 和 @donkeykong91。 - **PR #89736** 修复(status)：将低于 1000 的令牌计数渲染为普通整数。关联 #89735。感谢 @jbetala7 和 @vincentkoc。 - **PR #92792** 修复(agents)：在 sanitizeContentBlocksImages 中捕获格式错误的图像块。感谢 @LowCode191 和 @vincentkoc。 - **PR #92555** ci：在 Windows 配套资源上把关稳定版发布。感谢 @fuller-stack-dev。 - **PR #91824** 修复(agents)：在 sessions_spawn 工具描述中添加使用指南（修复 #91814）。感谢 @zenglingbiao 和 @cattails-lgao。 - **PR #92823** 修复(qqbot)：暴露失败的媒体发送。感谢 @zhangguiping-xydt。 - **PR #92045** 修复诊断 OTEL 运行时安装信任。感谢 @efpiva。 - **PR #91586** 修复(update)：在包医生警告后继续。感谢 @fuller-stack-dev。 - **PR #67783** 修复(feishu)：入站时定位输入反应。感谢 @huiwen01。 - **PR #73958** 修复(feishu)：保留 root_id 线程路由而不强制 thread_id。感谢 @huiwen01。 - **PR #68106** 修复(lobster)：暴露工作流路径错误。关联 #68101。感谢 @vvitovec 和 @MPC7500。 - **PR #90682** 修复(openai)：保留不透明的推理转录字段。关联 #90093。感谢 @toruvieI 和 @richardmqq。 - **PR #92373** 修复(anthropic)：当思考被禁用时从历史中剥离思考块（修复 #92360）。感谢 @liuhao1024 和 @notnaji。 - **PR #87346** 修复(anthropic)：在轮次验证中合并连续的助手轮次。关联 #87329。感谢 @Jefsky 和 @travellingsoldier85。 - **PR #92896** 修复(anthropic)：隔离无效的直接工具架构。 - **PR #89418** 修复(anthropic)：隔离无效的投影工具。感谢 @vincentkoc。 - **PR #89221** 修复(agents)：保护 Anthropic 工具描述符。感谢 @vincentkoc。 - **PR #90228** 修复(agents)：保护 Anthropic 工具架构转换。感谢 @vincentkoc。 - **PR #89622** 修复(agents)：跳过不可读的 Anthropic 工具架构。感谢 @vincentkoc。 - **PR #89229** 修复(llm)：保护 Anthropic 提供商工具描述符。感谢 @vincentkoc。 - **PR #90278** 修复(providers)：跳过不可读的 Anthropic 工具架构。感谢 @vincentkoc。 - **PR #90739** 修复(active-memory)：保留详细回忆摘要。感谢 @brokemac79。 - **PR #92558** 简化 QA 评分卡映射形状。感谢 @RomneyDa。 - **PR #92876** 修复(memory-wiki)：停止将原始源页面标记为格式错误。感谢 @vincentkoc。 - **PR #92908** 修复(providers)：隔离不可读的 Anthropic 载荷工具。 - **PR #92881** 修复(memory)：保留重建回滚恢复。感谢 @TSHOGX 和 @vincentkoc。 - **PR #92921** 修复(openai)：隔离不可读的工具架构。 - **PR #89413** 修复(openai)：隔离不可读的投影工具。感谢 @vincentkoc。 - **PR #89013** 修复(agents)：物化 OpenAI 工具架构。感谢 @vincentkoc。 - **PR #89016** 修复(agents)：保护 OpenAI 传输工具描述符。感谢 @vincentkoc。 - **PR #89378** 修复(agents)：保护 OpenAI 工具架构转换。感谢 @vincentkoc。 - **PR #89543** 修复(agents)：加强 OpenAI 严格架构检查。感谢 @vincentkoc。 - **PR #90200** 修复(agents)：保护 OpenAI 严格工具诊断。感谢 @vincentkoc。 - **PR #90283** 修复(providers)：跳过不可读的 OpenAI 补全工具架构。感谢 @vincentkoc。 - **PR #90286** 修复(providers)：跳过不可读的 OpenAI 响应工具架构。感谢 @vincentkoc。 - **PR #90397** 修复(openai)：跳过不可读的响应工具架构。感谢 @vincentkoc。 - **PR #92550** 将 Telegram RTT 采样纳入实时 QA 证据。感谢 @RomneyDa。 - **PR #92824** 修复(media)：通过 Codex 路由 OAuth 图像默认值。关联 #87168。感谢 @bek91。 - **PR #92911** 修复(cli)：避免最新标签的虚假降级提示。感谢 @Andy312432 和 @vincentkoc。 - **PR #92928** 修复(openai)：保护后钩子工具载荷。 - **PR #89703** 修复(openai)：保护响应工具载荷名称。感谢 @vincentkoc。 - **PR #92879** 修复(sessions)：在缺少异步转录时回退到重置存档。感谢 @masatohoshino 和 @CadanHu 和 @vincentkoc。 - **PR #92814** 修复(feishu)：当动态代理绑定已存在于运行时配置中时重新解析路由（修复 #42837）。感谢 @liuhao1024 和 @vincentkoc 和 @cwlong163-afk。 - **PR #90574** 修复(openai)：省略 gpt-5.5 工具推理努力。感谢 @BSG2000。 - **PR #92941** 修复(openai)：恢复无效的推理签名。 - **PR #92002** 修复(lmstudio)：向二进制思维模型发送思考“关闭”。感谢 @nxmxbbd。 - **PR #92916** #92201：嵌入式运行器：新流式传输的思维签名在重放时偶尔无效（Anthropic）；恢复包装器从未触发，因为错误文本已被泛化。感谢 @mmyzwl 和 @CarlCapital。 - **PR #90936** 修复(agents)：不要将客户端断开中止误分类为运行超时。感谢 @openperf。 - **PR #93009** 修复(agents)：使 wrapToolWithBeforeToolCallHook 幂等以防止双重钩子执行（修复 #92973）。感谢 @zenglingbiao 和 @dertbv。 - **PR #92318** 修复(cron)：需要显式消息目标证据。感谢 @hxy91819。 - **PR #93022** 修复(gateway)：修复跨代理的使用成本聚合。感谢 @luke-skywalker-open-claw 和 @stablegenius49。 - **PR #93159** 修复(tui)：在退出后保持父进程 stdin 暂停。感谢 @fuller-stack-dev。 - **PR #93616** 保持无密钥网页搜索提供商为选择加入。感谢 @davemorin 和 @vincentkoc。 - **PR #92679** feat(telegram)：发送富消息文本。感谢 @obviyus。 - **PR #93164** 修复(telegram)：保留富 markdown 换行。感谢 @vincentkoc。 - **PR #84082** 修复(telegram)：允许可展开引用块。感谢 @jzakirov。 - **PR #89421** 修复(telegram)：暴露线程创建 CLI 重映射。感谢 @spacegeologist。 - **PR #92788** 修复(sessions)：从发送策略中的账户范围 DM 会话密钥衍生渠道。感谢 @yetval。 - **PR #91246** 修复 WebChat 媒体补全交接。感谢 @TurboTheTurtle。 - **PR #92468** 修复 #92453：将会话身份添加到运行时提示。感谢 @zhangguiping-xydt 和 @QQSHI13。 - **PR #92657** feat(usage)：推出内置 /usage 完整页脚。感谢 @Marvinthebored。 - **PR #92779** 修复：工作区文件默认折叠。感谢 @shakkernerd。 - **PR #91353** 修复(ui)：在 Control UI 调度中保留 /reset 软参数。感谢 @zhouhe-xydt。 - **PR #92650** 修复 #92465：拆分 OpenAI 431 嵌入批次。感谢 @mushuiyu886 和 @BrettHamlin。 - **PR #91247** 修复(state)：在 NFS 状态卷上避免 SQLite WAL。感谢 @849261680。 - **PR #91218** 修复(google)：从 Vertex 模型路径中去除提供商前缀。感谢 @bymle。 - **PR #92162** #92069：修复(cli)：使用错误退出码 0。感谢 @Pandah97 和 @marcospaulo。 - **PR #92735** 修复(markdown-core)：将 Infinity 块限制视为无界，而非 1。感谢 @yhterrance。 - **PR #88748** 修复(gemini)：将 OAuth 配置文件桥接到 CLI 运行时。感谢 @jason-allen-oneal。 - **PR #64734** 修复(discord)：提高线程标题超时和令牌以适应推理模型。感谢 @hanamizuki。 - **PR #91287** 修复(cron)：在心跳模型输入中去重主会话 systemEvent。感谢 @ZengWen-DT。 - **PR #92510** 修复(gateway)：拒绝未知的 OpenAI 代理选择器。感谢 @zhangguiping-xydt。 - **PR #92646** 修复：需要管理员权限才能进行 HTTP 模型覆盖。感谢 @steipete-oai。 - **PR #90706** 修复(OpenAI Responses)：为无存储提供商禁用项目 ID 重放。感谢 @snowzlm。 - **PR #59137** 修复(memory)：在重建回滚中保留重试状态和嵌入缓存。感谢 @TSHOGX。 - **PR #90658** 修复(ui)：当会话列表过时时保留仪表盘会话父级谱系。感谢 @luoyanglang。 - **PR #92773** 修复(tui)：在 /model 确认中显示解析后的规范模型引用。感谢 @NarahariRaghava。 - **PR #92677** docs(nodes)：在 Nodes 概述中添加 openclaw.json 配置示例。感谢 @liuhao1024。 - **PR #92695** docs(config)：在 agent-defaults 类型注释中更正 maxConcurrent 默认值（AI 辅助）。感谢 @ArielSmoliar。 ### 发布验证 - npm 包：https://www.npmjs.com/package/openclaw/v/2026.6.8 - 注册表压缩包：https://registry.npmjs.org/openclaw/-/openclaw-2026.6.8.tgz - 完整性校验：`sha512-iziR8fi69+ojrtX7FYYvTpkGcVnmyLpIhvchgt5LFkkdHVWw973XAAekKVZ3/xQJ5FG4NwgHkXL0LLTrgsNOSQ==` - 发布 SHA：`844f405ac1be805d5c598922a37254f12ab6d765` - 完整发布 CI 报告：https://github.com/openclaw/releases/blob/main/evidence/2026.6.8/release-evidence.md - 发布发布：https://github.com/openclaw/openclaw/actions/runs/27631170936 - npm 预检：https://github.com/openclaw/openclaw/actions/runs/27627935264 - 完整发布验证：https://github.com/openclaw/openclaw/actions/runs/27627935341 - 插件 npm 发布：https://github.com/openclaw/openclaw/actions/runs/27631412538 - 插件 ClawHub 发布：https://github.com/openclaw/openclaw/actions/runs/27631418543 - 插件 ClawHub 引导：不需要 - OpenClaw npm 发布：https://github.com/openclaw/openclaw/actions/runs/27632229765 - npm Telegram beta E2E：未提供 - Windows Hub 升级：https://github.com/openclaw/openclaw/actions/runs/27632639125 来自 openclaw/[email protected] - macOS 签名/公证预检：https://github.com/openclaw/releases/actions/runs/27632793120 - macOS Swift 验证：https://github.com/openclaw/releases/actions/runs/27632792831 - macOS 资源升级：https://github.com/openclaw/releases/actions/runs/27635471007 - 稳定版 appcast：https://github.com/openclaw/openclaw/pull/93722

## 2026.6.8 ### Highlights - **Richer channel delivery:** Telegram and WhatsApp are less brittle: Telegram renders structured text with tables, lists, expandable blockquotes, preserved intentional line breaks, and CLI-backed replies, while WhatsApp now honors configured ACP bindings. (#92679, #93164, #84082, #89421, #92513) Thanks @obviyus, @vincentkoc, @jzakirov, @spacegeologist, @TurboTheTurtle, and @mcaxtr. - **More reliable agent runs:** account-scoped DM sends, generated media completions, auto-reply message-tool final replies, reset archive fallback reads, restart shutdown aborts, yielded subagent pauses, and session identity prompts all stay on the correct recovery path. (#92788, #91246, #92879, #91357, #92631, #92468) Thanks @yetval, @TurboTheTurtle, @masatohoshino, @CadanHu, @vincentkoc, @ooiuuii, @openperf, @zhangguiping-xydt, and @QQSHI13. - **Safer model routing:** new GLM-5.2 and Claude Haiku 4.5 catalog support arrives with normalized provider IDs, managed SecretRef auth, bounded model browsing, and safer OpenAI/Anthropic tool-schema recovery. (#92796, #90116, #92627, #90686, #92247, #92941) Thanks @arkyu2077, @liuhao1024, @lijenhsin, @rohitjavvadi, and @samson910022. - **Useful usage footers:** `/usage` and reply payload hooks now have a native full footer renderer, default template, fixed-decimal formatting, credential-aware limits, better partial-count handling, and warnings for broken templates instead of silent bad output. (#92657, #89835, #89629) Thanks @Marvinthebored. - **Predictable web search defaults:** key-free providers such as Parallel Free, DuckDuckGo, Ollama, and Codex Hosted Search remain explicit opt-ins rather than surprising automatic fallbacks. (#93616) Thanks @davemorin and @vincentkoc. - **Calmer UI and mobile sessions:** workspace files start collapsed, WebChat backscroll survives streaming, the desktop session picker remains interactive, reset arguments survive dispatch, and iOS reconnects stale foreground Gateways. (#92779, #92622, #92705, #91353, #92552) Thanks @shakkernerd, @TurboTheTurtle, @NianJiuZst, @zhouhe-xydt, and @Solvely-Colin. - **Resilient memory and state:** oversized OpenAI embedding batches split before 431s, QMD search stays available in transient mode, SQLite avoids WAL on NFS volumes, and full reindexes preserve rollback/cache recovery. (#92650, #92618, #92639, #91247, #92881) Thanks @mushuiyu886, @BrettHamlin, @zhbcher, @TurboTheTurtle, @Takhoffman, @849261680, and @TSHOGX. ### Changes - Providers/models: add GLM-5.2 support and Claude Haiku 4.5 catalog entries while keeping provider-qualified model IDs normalized across OpenRouter and Google Vertex paths. (#92796, #90116, #92627, #91218) Thanks @arkyu2077, @liuhao1024, and @bymle. - Web search: keep key-free providers such as Parallel Free, DuckDuckGo, Ollama, and Codex Hosted Search as explicit opt-ins instead of selecting them automatically when no API-backed provider is configured. (#93616) Thanks @davemorin and @vincentkoc. - Channel plugins: ship Telegram rich-message delivery and WhatsApp ACP binding support, including preserved intentional line breaks, rich prompt handoff to CLI backends, and transport fixtures for richer drafts. (#92679, #93164, #92513) Thanks @obviyus and @TurboTheTurtle. - Agent commands: support `/btw` in CLI-backed sessions and keep CLI usage-error exits classified as usage failures instead of successful runs. (#92669, #92162) Thanks @joshavant and @Pandah97. - Usage hooks: add built-in full footer rendering, default footer templates, per-turn usage state, credential-aware limits, and fixed-decimal formatting for usage-bar templates. (#92657, #89835, #89629) Thanks @Marvinthebored. ### Fixes - Channels and delivery: preserve account-scoped DM channel send policy, intentional rich-message line breaks in Telegram and status output, rich Telegram final replies, rich Telegram tables and lists, Telegram thread-create CLI remapping, Feishu dynamic-agent routes after persisted binding reuse, Slack outbound `message_sent` hooks, contributed message-tool schema optionality, same-channel generated media completions, and channel chunking around surrogate pairs and Infinity limits. (#92788, #93164, #92679, #89421, #89943, #42837, #92814, #91137, #91246, #92735) Thanks @yetval, @obviyus, @spacegeologist, @rishitamrakar, @liuhao1024, @lundog, @TurboTheTurtle, and @yhterrance. - Gemini CLI: use the selected OpenClaw OAuth/API-key auth profile in an isolated Gemini CLI runtime home, preventing ambient Google machine credentials from overriding the chosen profile. (#88748) Thanks @jason-allen-oneal and @shakkernerd. - Discord: give generated auto-thread titles a 60-second timeout and 4,096-token reasoning-model output budget, clamped to the selected model output cap. (#64734) Thanks @hanamizuki. - Agent, cron, and Gateway runtime: mark active main sessions before restart shutdown aborts, pause yielded subagent runs whose terminal also signals abort, clamp trusted subagent thinking overrides through provider/model fallback, preserve yielded media completions, deliver channel message-tool final replies through auto-reply while hiding internal delivery hints, restore reset archive fallback reads when active async transcripts are missing, de-duplicate main-session heartbeat events, expose session identity in runtime prompts, reject unknown OpenAI agent selectors, keep generated media completions, slash-command block replies, and trajectory export commands in WebChat, and require admin privileges for HTTP session/model override surfaces. (#91357, #92631, #92412, #92146, #92879, #91287, #92468, #92510, #91246, #92651, #92646) Thanks @ooiuuii, @openperf, @IWhatsskill, @masatohoshino, @CadanHu, @ZengWen-DT, @zhangguiping-xydt, @TurboTheTurtle, and @oiGaDio. - Providers and model replay: preserve storeless OpenAI Responses replay compatibility, recover invalid OpenAI reasoning signatures and genericized Anthropic thinking-signature replay errors, route OAuth image defaults through Codex for eligible OpenAI profiles, avoid eager tool streaming for Claude 4.5 in Copilot, quarantine unreadable and post-hook OpenAI/Anthropic-family tool schemas without broadening allowed tool choices, deliver explicit thinking-off requests to LM Studio binary-thinking models, honor profile auth for SecretRef model entries, bound model browsing, strip provider prefixes where runtimes need bare IDs, and surface nested embedding fetch failures. (#90706, #92941, #92201, #92916, #92824, #75393, #92908, #92921, #92928, #92002, #90686, #92247, #92627, #91218, #92628) Thanks @snowzlm, @mmyzwl, @CarlCapital, @bek91, @Kailigithub, @vincentkoc, @rohitjavvadi, @samson910022, @nxmxbbd, @liuhao1024, @bymle, and @mushuiyu886. - Memory, state, diagnostics, and config: split header-too-large embedding batches, keep QMD memory search enabled in transient mode, avoid SQLite WAL on NFS volumes, preserve recovery scheduling outside stuck-session warning backoff, preserve full-reindex rollback/cache recovery, and treat raw Memory Wiki source pages as source evidence. (#92650, #92618, #92639, #91247, #92752, #92881, #59137, #92876) Thanks @mushuiyu886, @TurboTheTurtle, @849261680, @gnanam1990, @TSHOGX, and @arlen8411. - UI/mobile/TUI: preserve dashboard session parent lineage, WebChat backscroll, reset soft command args, sidebar session picker interactivity, collapsed workspace files, resolved `/model` confirmation refs, stale foreground iOS Gateway reconnects, and paused setup-parent stdin after inherited-stdio child exit. (#90658, #92622, #91353, #92705, #92779, #92773, #92552, #93159) Thanks @luoyanglang, @TurboTheTurtle, @zhouhe-xydt, @NianJiuZst, @shakkernerd, @NarahariRaghava, @Solvely-Colin, and @fuller-stack-dev. - Plugins and updates: repair missing required platform packages during managed plugin installs and updates, including omitted Codex platform binaries. - Dependencies: update Hono to 4.12.25 so published OpenClaw and ACPX packages use the patched runtime. - Release reliability: extend slow Gateway/full-suite watchdogs, split local full-suite shards when throttled, fold Telegram RTT sampling into live QA evidence, simplify QA scorecard mappings around canonical coverage IDs, preserve release-check child refs while pinning expected SHAs, widen live OpenAI TTS budgets for slower provider responses, and avoid false downgrade prompts for unresolved latest-tag updates. (#92550, #92558, #92911) Thanks @RomneyDa and @Andy312432. ### Complete contribution record This audited record covers the complete v2026.6.6..v2026.6.8 history: 168 merged PRs. The generation manifest also supplies direct commits as editorial input; the grouped notes above prioritize user impact. #### Pull requests - **PR #92144** fix(cron): report SQLite storage path in cron.status instead of legacy jobs.json. Related #91766. Thanks @liuhao1024 and @AaronFaby. - **PR #92175** fix(channel): harden local setup trust. Thanks @hxy91819. - **PR #92073** fix: handle explicit silent assistant replies. Thanks @sallyom. - **PR #92087** fix(docker): bundle QA Lab runtime in the image. Thanks @jesse-merhi. - **PR #92387** fix(anthropic-vertex): stop re-marking cache_control on transport-budgeted payloads. Thanks @openperf and @Takhoffman. - **PR #92229** Fix doctor preview channel SecretRef resolution. Thanks @joshavant. - **PR #92225** Fix disabled heartbeat one-shot cron retries. Thanks @joshavant. - **PR #92265** Fix configured DeepSeek model transport inheritance. Thanks @joshavant. - **PR #92226** Fail closed for CLI-backed /btw fallback. Thanks @joshavant. - **PR #92231** Fix suppressed heartbeat commitment delivery. Thanks @joshavant. - **PR #92280** fix(agents): classify structured unsupported model errors. Thanks @joshavant. - **PR #92276** Fix OTLP log trace correlation. Thanks @joshavant. - **PR #92282** fix(update): hand off Linux service auto-updates. Thanks @joshavant. - **PR #92235** fix: resolve managed SecretRef provider auth. Thanks @joshavant. - **PR #92293** Fix provider static model fallback resolution. Thanks @joshavant. - **PR #92343** fix(agent): continue after source message tool replies. Thanks @joshavant. - **PR #92350** fix(codex): preserve memory prompt registration. Thanks @rubencu. - **PR #92290** fix: clarify gateway SecretRef auth diagnostics. Thanks @joshavant. - **PR #92286** fix: repair rejected Anthropic thinking replay. Thanks @joshavant. - **PR #92281** Fix Telegram spooled buffered replay. Thanks @joshavant. - **PR #92416** fix(outbound): honor top-level image param as send media source (#92407). Thanks @xydigit-sj and @ichirokyoto. - **PR #92508** fix(sandbox): render CLI skill prompts from materialized paths. Thanks @brokemac79. - **PR #92540** chore: fix esbuild production audit failure. Thanks @RomneyDa. - **PR #91484** Add QA evidence artifact output. Thanks @RomneyDa. - **PR #91500** Add QA scorecard taxonomy validation. Thanks @RomneyDa. - **PR #92554** feat(moonshot): add Kimi K2.7 Code support. - **PR #92396** fix(moonshot): backfill reasoning_content on assistant tool-call replay messages. Related #71491. Thanks @xialonglee and @RoseKongPS. - **PR #92566** Fix lifecycle timeout cleanup after leader exit. Thanks @RomneyDa. - **PR #88993** Expose paged channel action results. Thanks @fuller-stack-dev. - **PR #90326** fix(fireworks): resolve catalog model params from plugin.json via core. Thanks @obuchowski. - **PR #86629** fix(doctor): warn for untrusted external Discord plugin. Related #83212. Thanks @brokemac79 and @cdeyoung67. - **PR #90242** fix(providers): skip unreadable Mistral tool schemas. Thanks @vincentkoc. - **PR #92498** fix(reply): mirror same-channel Slack final replies. Related #92489. Thanks @TurboTheTurtle and @TalkingHeadsJed. - **PR #92083** fix(channels): default boundary logger for swallowed progress-draft start errors. Thanks @hansraj316. - **PR #92031** fix(channels): make timer-fired progress-draft start errors observable. Thanks @hansraj316. - **PR #92564** fix(agents): isolate invalid plugin model catalogs [AI-assisted]. Related #92553. Thanks @tangtaizong666 and @fxstein. - **PR #89827** docs: UX-013 — design system documentation. Thanks @BunsDev. - **PR #89615** feat(ui): hide empty workboard columns. Thanks @BunsDev. - **PR #89822** fix(a11y): B-1+B-2+B-3 — contrast, focus states, minimum font sizes. Thanks @BunsDev. - **PR #92618** fix #92218: memory_search tool disabled with QMD backend. Thanks @mushuiyu886 and @zhbcher. - **PR #92608** docs(gateway): add uptime monitoring guidance to health check docs (fixes #55768). Thanks @liuhao1024 and @faahim. - **PR #92605** fix(docs): pin Windows Hub download links to v2026.6.5. Related #92470. Thanks @lzyyzznl and @arjkul. - **PR #92593** #92589: fix(internal-runtime-context): wrap prompt-preface runtime context body in delimiters. Thanks @zhangqueping and @jovi2014-cyber. - **PR #92606** Run Vitest and Playwright scenarios from qa suite. Thanks @RomneyDa. - **PR #89629** feat(hooks): per-turn usageState on reply_payload_sending. Thanks @Marvinthebored. - **PR #89835** feat(usage): native templated /usage full footer renderer. Thanks @Marvinthebored. - **PR #92247** fix(models): bound /models and models list catalog loading. Related #91809. Thanks @samson910022 and @syfvb. - **PR #90686** fix(gateway): honor profile auth for SecretRef model entries. Related #90685. Thanks @rohitjavvadi. - **PR #92651** fix: require admin for HTTP session kills. Thanks @steipete-oai. - **PR #92652** test(models): stabilize plugin auth marker fixtures. - **PR #89438** fix(slack): warn when channels map is keyed by name instead of channel ID. Related #81665. Thanks @Alix-007 and @cjalden. - **PR #92631** fix(agents): pause yielded subagent runs whose terminal also signals abort. Related #92448. Thanks @openperf and @aleps001. - **PR #92622** fix(ui): preserve WebChat backscroll during streaming. Thanks @TurboTheTurtle. - **PR #92627** fix(openrouter): strip openrouter/ prefix from model ID in normalizeResolvedModel hook (fixes #92611). Thanks @liuhao1024 and @lijenhsin. - **PR #92146** fix(cron): preserve yielded media completions. Thanks @IWhatsskill. - **PR #90116** fix: add Claude Haiku 4.5 static catalog entries. Related #90088. Thanks @arkyu2077 and @maaron34. - **PR #91137** fix(channels): keep contributed message-tool schema properties optional. Thanks @lundog. - **PR #75393** fix(copilot): disable eager tool streaming for Claude 4.5. Related #75348. Thanks @Kailigithub and @finchinslc. - **PR #92628** fix #73713: surface nested embedding fetch failures. Thanks @mushuiyu886 and @crsnpalmer-art. - **PR #89943** fix(slack): emit message_sent hook on outbound delivery (mirror Telegram). Related #89942. Thanks @rishitamrakar. - **PR #92668** fix(docs): finalize i18n postprocess before skip. Thanks @hxy91819. - **PR #92673** fix: split image setup and request timeout semantics. Thanks @hxy91819. - **PR #92639** fix(memory): keep memory_search in transient qmd mode. Thanks @TurboTheTurtle and @Takhoffman. - **PR #92705** fix(ui): restore sidebar session picker interactivity above desktop workbench. Thanks @NianJiuZst. - **PR #92669** feat: support /btw in CLI-backed sessions. Thanks @joshavant. - **PR #91357** fix(gateway): mark active main sessions before restart shutdown aborts. Thanks @ooiuuii. - **PR #92552** fix(ios): force stale foreground gateway reconnects. Thanks @Solvely-Colin. - **PR #92752** fix(diagnostics): keep recovery scheduling out of the stuck-session warning backoff. Thanks @gnanam1990 and @Takhoffman. - **PR #92766** clarify before_install hook scope. Thanks @sallyom. - **PR #92513** Honor WhatsApp configured ACP bindings. Thanks @TurboTheTurtle and @mcaxtr. - **PR #92796** feat(providers): add GLM-5.2 support. - **PR #92807** fix(heartbeat): route outbound mirror to isolated session key. Thanks @agent-merkava. - **PR #92745** fix(memory): explain skipped short-term recall hits. Thanks @mushuiyu886. - **PR #92488** fix(gateway): forward image-only input on /v1/responses (parity with chat completions). Thanks @s554097550. - **PR #92604** fix(status): avoid cumulative usage for context percent. Thanks @ashishpatel26. - **PR #92547** fix(nodes): surface pending reapproval diagnostics. Thanks @fuller-stack-dev. - **PR #92690** fix(doctor): avoid false-positive legacy cron store warning when store was already migrated (fixes #92683). Thanks @liuhao1024 and @motteman. - **PR #92806** fix(telegram): skip IPv4 fallback when user explicitly configures non-ipv4first dnsResultOrder (fixes #41671). Thanks @liuhao1024 and @vincentkoc and @leandroirani933-ctrl. - **PR #92778** fix(macos): defer isOverflowing mutation to break SwiftUI render loop (fixes #43480). Thanks @liuhao1024 and @vincentkoc and @gdiab. - **PR #92795** fix(gateway): use resolveNonNegativeNumber for totalTokens to display 0 instead of ? (fixes #43009). Thanks @liuhao1024 and @vincentkoc and @ltxy12138-ai. - **PR #92746** fix(gateway): preserve active runs during plugin finalization. Thanks @scotthuang and @vincentkoc. - **PR #92820** UI: localize Logs tab labels. Thanks @rubensfox20. - **PR #61080** UI: localize logs hardcoded labels. Thanks @rubensfox20. - **PR #92825** fix(telegram): preserve command callbacks while prefixing generic callback data. Related #54909. Thanks @hnshah and @timt80. - **PR #54962** fix(telegram): add 'callback_data:' prefix to inline button callbacks. Related #54909. Thanks @hnshah and @timt80. - **PR #87060** fix(copilot): drop thinking blocks from replay. Related #81520. Thanks @giodl73-repo and @warcold. - **PR #81534** fix(github-copilot): strip thinking blocks from latest assistant turn (#81520). Thanks @SymbolStar and @warcold. - **PR #92834** feat(browser): extend --labels overlay to full-page and element captures. Thanks @hxy91819. - **PR #92095** fix #92039: [Bug]: WhatsApp login reports success before auth is durably persisted, so Docker rebuilds/upgrades can force relink. Thanks @zhangguiping-xydt and @dinorastoder. - **PR #92801** fix(stale): exempt ClawSweeper actionable labels from stale lifecycle (fixes #89564). Related #78640, #81078, #81122. Thanks @liuhao1024 and @brokemac79 and @MilleniumGenAI and @Jackten and @donkeykong91. - **PR #89736** fix(status): render sub-1000 token counts as plain integers. Related #89735. Thanks @jbetala7 and @vincentkoc. - **PR #92792** fix(agents): catch malformed image blocks in sanitizeContentBlocksImages. Thanks @LowCode191 and @vincentkoc. - **PR #92555** ci: gate stable releases on Windows companion assets. Thanks @fuller-stack-dev. - **PR #91824** fix(agents): add usage guidance to sessions_spawn tool description (fixes #91814). Thanks @zenglingbiao and @cattails-lgao. - **PR #92823** fix(qqbot): surface failed media sends. Thanks @zhangguiping-xydt. - **PR #92045** Fix diagnostics OTEL runtime install trust. Thanks @efpiva. - **PR #91586** fix(update): continue after package doctor warnings. Thanks @fuller-stack-dev. - **PR #67783** fix(feishu): target typing reaction on inbound. Thanks @huiwen01. - **PR #73958** fix(feishu): preserve root_id thread routing without thread_id forcing. Thanks @huiwen01. - **PR #68106** fix(lobster): surface workflow path errors. Related #68101. Thanks @vvitovec and @MPC7500. - **PR #90682** fix(openai): preserve opaque reasoning transcript fields. Related #90093. Thanks @toruvieI and @richardmqq. - **PR #92373** fix(anthropic): strip thinking blocks from history when thinking is disabled (fixes #92360). Thanks @liuhao1024 and @notnaji. - **PR #87346** fix(anthropic): merge consecutive assistant turns in turn validation. Related #87329. Thanks @Jefsky and @travellingsoldier85. - **PR #92896** fix(anthropic): quarantine invalid direct tool schemas. - **PR #89418** fix(anthropic): quarantine invalid projected tools. Thanks @vincentkoc. - **PR #89221** fix(agents): guard Anthropic tool descriptors. Thanks @vincentkoc. - **PR #90228** fix(agents): guard Anthropic tool schema conversion. Thanks @vincentkoc. - **PR #89622** fix(agents): skip unreadable Anthropic tool schemas. Thanks @vincentkoc. - **PR #89229** fix(llm): guard Anthropic provider tool descriptors. Thanks @vincentkoc. - **PR #90278** fix(providers): skip unreadable Anthropic tool schemas. Thanks @vincentkoc. - **PR #90739** fix(active-memory): preserve verbose recall summaries. Thanks @brokemac79. - **PR #92558** Simplify QA scorecard mapping shape. Thanks @RomneyDa. - **PR #92876** fix(memory-wiki): stop flagging raw source pages as malformed. Thanks @vincentkoc. - **PR #92908** fix(providers): quarantine unreadable Anthropic payload tools. - **PR #92881** fix(memory): preserve reindex rollback recovery. Thanks @TSHOGX and @vincentkoc. - **PR #92921** fix(openai): quarantine unreadable tool schemas. - **PR #89413** fix(openai): quarantine unreadable projected tools. Thanks @vincentkoc. - **PR #89013** fix(agents): materialize OpenAI tool schemas. Thanks @vincentkoc. - **PR #89016** fix(agents): guard OpenAI transport tool descriptors. Thanks @vincentkoc. - **PR #89378** fix(agents): guard OpenAI tool schema conversion. Thanks @vincentkoc. - **PR #89543** fix(agents): harden OpenAI strict schema inspection. Thanks @vincentkoc. - **PR #90200** fix(agents): guard OpenAI strict tool diagnostics. Thanks @vincentkoc. - **PR #90283** fix(providers): skip unreadable OpenAI completion tool schemas. Thanks @vincentkoc. - **PR #90286** fix(providers): skip unreadable OpenAI responses tool schemas. Thanks @vincentkoc. - **PR #90397** fix(openai): skip unreadable responses tool schemas. Thanks @vincentkoc. - **PR #92550** Fold Telegram RTT sampling into live QA evidence. Thanks @RomneyDa. - **PR #92824** fix(media): route OAuth image defaults through Codex. Related #87168. Thanks @bek91. - **PR #92911** fix(cli): avoid false downgrade prompt for latest tag. Thanks @Andy312432 and @vincentkoc. - **PR #92928** fix(openai): guard post-hook tool payloads. - **PR #89703** fix(openai): guard responses tool payload names. Thanks @vincentkoc. - **PR #92879** fix(sessions): fall back to reset archive for missing async transcripts. Thanks @masatohoshino and @CadanHu and @vincentkoc. - **PR #92814** fix(feishu): re-resolve route when dynamic agent binding already exists in runtime config (fixes #42837). Thanks @liuhao1024 and @vincentkoc and @cwlong163-afk. - **PR #90574** fix(openai): omit gpt-5.5 tool reasoning effort. Thanks @BSG2000. - **PR #92941** fix(openai): recover invalid reasoning signatures. - **PR #92002** fix(lmstudio): deliver thinking "off" to binary-thinking models. Thanks @nxmxbbd. - **PR #92916** #92201: Embedded runner: freshly streamed thinking signatures intermittently invalid on replay (Anthropic); recovery wrapper never fires because error text is genericized. Thanks @mmyzwl and @CarlCapital. - **PR #90936** fix(agents): do not misclassify client-disconnect abort as run timeout. Thanks @openperf. - **PR #93009** fix(agents): make wrapToolWithBeforeToolCallHook idempotent to prevent double hook execution (fixes #92973). Thanks @zenglingbiao and @dertbv. - **PR #92318** fix(cron): require explicit message target proof. Thanks @hxy91819. - **PR #93022** fix(gateway): repair usage cost aggregation across agents. Thanks @luke-skywalker-open-claw and @stablegenius49. - **PR #93159** fix(tui): keep parent stdin paused after exit. Thanks @fuller-stack-dev. - **PR #93616** Keep key-free web search providers opt-in. Thanks @davemorin and @vincentkoc. - **PR #92679** feat(telegram): send rich message text. Thanks @obviyus. - **PR #93164** fix(telegram): preserve rich markdown line breaks. Thanks @vincentkoc. - **PR #84082** fix(telegram): allow expandable blockquotes. Thanks @jzakirov. - **PR #89421** fix(telegram): expose thread create CLI remap. Thanks @spacegeologist. - **PR #92788** fix(sessions): derive channel from account-scoped DM session keys in send-policy. Thanks @yetval. - **PR #91246** Fix webchat media completion handoff. Thanks @TurboTheTurtle. - **PR #92468** fix #92453: add session identity to runtime prompt. Thanks @zhangguiping-xydt and @QQSHI13. - **PR #92657** feat(usage): ship built-in /usage full footer. Thanks @Marvinthebored. - **PR #92779** fix: start workspace files collapsed. Thanks @shakkernerd. - **PR #91353** fix(ui): preserve /reset soft args in Control UI dispatch. Thanks @zhouhe-xydt. - **PR #92650** fix #92465: split OpenAI 431 embedding batches. Thanks @mushuiyu886 and @BrettHamlin. - **PR #91247** fix(state): avoid sqlite wal on nfs state volumes. Thanks @849261680. - **PR #91218** fix(google): strip provider prefix from Vertex model path. Thanks @bymle. - **PR #92162** #92069: fix(cli): usage errors exit 0. Thanks @Pandah97 and @marcospaulo. - **PR #92735** fix(markdown-core): treat Infinity chunk limit as unbounded, not 1. Thanks @yhterrance. - **PR #88748** fix(gemini): bridge OAuth profiles into CLI runtime. Thanks @jason-allen-oneal. - **PR #64734** fix(discord): raise thread title timeout and tokens to fit reasoning models. Thanks @hanamizuki. - **PR #91287** fix(cron): de-duplicate main-session systemEvent in heartbeat model input. Thanks @ZengWen-DT. - **PR #92510** fix(gateway): reject unknown OpenAI agent selectors. Thanks @zhangguiping-xydt. - **PR #92646** fix: require admin for HTTP model overrides. Thanks @steipete-oai. - **PR #90706** fix(OpenAI Responses): disable item id replay for storeless providers. Thanks @snowzlm. - **PR #59137** fix(memory): preserve retry state and embedding cache across reindex rollback. Thanks @TSHOGX. - **PR #90658** fix(ui): preserve dashboard session parent lineage when session list is stale. Thanks @luoyanglang. - **PR #92773** fix(tui): show resolved canonical model ref in /model confirmation. Thanks @NarahariRaghava. - **PR #92677** docs(nodes): add openclaw.json config example to Nodes overview. Thanks @liuhao1024. - **PR #92695** docs(config): correct maxConcurrent default in agent-defaults type comments (AI-assisted). Thanks @ArielSmoliar. ### Release verification - npm package: https://www.npmjs.com/package/openclaw/v/2026.6.8 - registry tarball: https://registry.npmjs.org/openclaw/-/openclaw-2026.6.8.tgz - integrity: `sha512-iziR8fi69+ojrtX7FYYvTpkGcVnmyLpIhvchgt5LFkkdHVWw973XAAekKVZ3/xQJ5FG4NwgHkXL0LLTrgsNOSQ==` - release SHA: `844f405ac1be805d5c598922a37254f12ab6d765` - full release CI report: https://github.com/openclaw/releases/blob/main/evidence/2026.6.8/release-evidence.md - release publish: https://github.com/openclaw/openclaw/actions/runs/27631170936 - npm preflight: https://github.com/openclaw/openclaw/actions/runs/27627935264 - full release validation: https://github.com/openclaw/openclaw/actions/runs/27627935341 - plugin npm publish: https://github.com/openclaw/openclaw/actions/runs/27631412538 - plugin ClawHub publish: https://github.com/openclaw/openclaw/actions/runs/27631418543 - plugin ClawHub bootstrap: not needed - OpenClaw npm publish: https://github.com/openclaw/openclaw/actions/runs/27632229765 - npm Telegram beta E2E: not supplied - Windows Hub promotion: https://github.com/openclaw/openclaw/actions/runs/27632639125 from openclaw/[email protected] - macOS signed/notarized preflight: https://github.com/openclaw/releases/actions/runs/27632793120 - macOS Swift validation: https://github.com/openclaw/releases/actions/runs/27632792831 - macOS asset promotion: https://github.com/openclaw/releases/actions/runs/27635471007 - stable appcast: https://github.com/openclaw/openclaw/pull/93722