OpenClaw v2026.4.5 - 更新日志

### 破坏性变更 - 配置：移除遗留公共配置别名，例如 `talk.voiceId` / `talk.apiKey`、`agents.*.sandbox.perSession`、`browser.ssrfPolicy.allowPrivateNetwork`、`hooks.internal.handlers` 以及频道/群组/房间的 `allow` 切换开关，转而使用标准公共路径和 `enabled` 字段，同时保持对现有配置的加载时兼容性，并支持通过 `openclaw doctor --fix` 进行迁移。（#60726）感谢 @vincentkoc。 ### 变更 - 智能体/视频生成：添加内置的 `video_generate` 工具，使智能体能够通过配置的提供商创建视频，并在回复中直接返回生成的媒体。 - 智能体/音乐生成：忽略不支持的可选提示（例如 `durationSeconds`），仅发出警告而非在 Google Lyria 等提供商处直接导致请求失败。 - 提供商/ComfyUI：为本地 ComfyUI 和 Comfy Cloud 工作流添加捆绑的 `comfy` 工作流媒体插件，包括共享的 `image_generate`、`video_generate` 以及基于工作流的 `music_generate` 支持，具备提示注入、可选参考图片上传、实时测试和输出下载功能。 - 工具/音乐生成：添加内置的 `music_generate` 工具，捆绑 Google (Lyria) 和 MiniMax 提供商，并支持基于工作流的 Comfy 集成，包括异步任务跟踪和完成后音频的后续传递。 - 提供商：添加捆绑的 Qwen、Fireworks AI 和 StepFun 提供商，以及用于聊天、语音和搜索工作流的 MiniMax TTS、Ollama Web Search 和 MiniMax Search 集成。（#60032, #55921, #59318, #54648） - 提供商/Amazon Bedrock：添加捆绑的 Mantle 支持，以及推理配置文件发现和自动请求区域注入，使 Bedrock 托管的 Claude、GPT-OSS、Qwen、Kimi、GLM 及类似路由所需的手动设置更少。（#61296, #61299）感谢 @wirjo。 - 控制 UI/多语言：为简体中文、繁体中文、巴西葡萄牙语、德语、西班牙语、日语、韩语、法语、土耳其语、印尼语、波兰语和乌克兰语添加本地化的控制 UI 支持。感谢 @vincentkoc。 - 插件：在引导式入门/设置流程中添加插件配置 TUI 提示，并添加 `openclaw plugins install --force` 以便在不使用危险代码覆盖标志的情况下替换现有插件和钩子包目标。（#60590, #60544） - 控制 UI/技能：在技能面板中直接添加 ClawHub 搜索、详情和安装流程。（#60134）感谢 @samzong。 - iOS/执行审批：添加通用的 APNs 审批通知，打开应用内执行审批模态框，仅在经过身份验证的操作员重新连接后获取命令详情，并在审批解决时清除过期的通知状态。（#60239）感谢 @ngutman。 - Matrix/执行审批：添加原生 Matrix 执行审批提示，支持账户范围的审批者、频道或私信传递，以及感知房间线程的解析处理。（#58635）感谢 @gumadeiras。 - 频道/上下文可见性：为每个频道添加可配置的 `contextVisibility`（`all`、`allowlist`、`allowlist_quote`），使补充引用、线程和获取的历史上下文能够按发送者允许列表进行过滤，而非始终按接收时传递。 - 提供商/请求覆盖：在 OpenAI、Anthropic、Google 及兼容提供商路径之间添加共享的模型和媒体请求传输覆盖，包括标头、认证、代理和 TLS 控制。（#60200） - 提供商/OpenAI：添加前向兼容的 `openai-codex/gpt-5.4-mini`、一个可选入的 GPT 人格，以及提供商拥有的 GPT-5 提示贡献，使 Codex/GPT 运行保持缓存稳定，并与捆绑目录滞后兼容。 - 智能体/Claude CLI：通过环回 MCP 桥向后台 Claude CLI 运行公开 OpenClaw 工具，并将捆绑运行切换到 stdin + `stream-json` 部分消息流式传输，以便提示停止通过 argv 传递、长回复显示实时进度，且最终会话/使用元数据仍能干净落地。（#35676）感谢 @mylukin。 - ACPX/运行时：将 ACP 运行时直接嵌入捆绑的 `acpx` 插件，移除额外的外部 ACP CLI 中间环节，强化实时 ACP 会话绑定和重用，并添加通用的 `reply_dispatch` 钩子，以便 ACPX 等捆绑插件能在无需核心自动回复路由中硬编码 ACP 路径的情况下拥有回复拦截权。（#61319） - 智能体/进度：添加实验性的结构化计划更新和结构化执行项事件，以便兼容的 UI 能在长时间运行期间显示更清晰的逐步进度。 - 提供商/Anthropic：在新引导流程中移除 Claude CLI 后端和 setup-token，保持现有配置的遗留配置文件可运行，并在迁移期间让 `openclaw doctor` 修复或移除过时的 `anthropic:claude-cli` 状态。 - 工具/视频生成：添加捆绑的 xAI (`grok-imagine-video`)、阿里巴巴 Model Studio Wan 和 Runway 视频提供商，并为这三者添加实时测试/默认模型连接。 - 内存/搜索：为 Titan、Cohere、Nova 和 TwelveLabs 模型添加 Amazon Bedrock 嵌入，支持 `provider: "auto"` 的 AWS 凭证链自动检测以及提供商特定的维度控制。感谢 @wirjo。 - 提供商/Amazon Bedrock Mantle：从 AWS 凭证链生成持有者令牌，使 Mantle 自动发现能够使用 IAM 认证，而无需手动导出 `AWS_BEARER_TOKEN_BEDROCK`。感谢 @wirjo。 - 内存/做梦（实验性）：添加加权的短期回忆提升、`/dreaming` 命令、Dreams UI、多语言概念标记以及 doctor/status 修复支持，同时将做梦从竞争模式重构为三个协作阶段（浅睡、深睡、REM），具有独立调度和恢复行为，以便持久记忆提升能在后台运行，减少手动设置。（#60569, #60697）感谢 @vignesh07。 - 内存/做梦：添加可配置的衰老控制（`recencyHalfLifeDays`、`maxAgeDays`）以及可选详细日志记录，以便操作员可以调整回忆衰减并更轻松地检查提升决策。 - 内存/做梦：添加 REM 预览工具（`openclaw memory rem-harness`、`promote-explain`），在 REM 阶段揭示可能的持久真相，并使深度提升可重放安全，以便重运行能协调而非重复 `MEMORY.md` 条目。 - 内存/做梦：将做梦轨迹内容写入顶层 `dreams.md` 而非每日记忆笔记，更新 `/dreaming` 帮助文本指向该文件，并保持 `dreams.md` 可用于显式读取，而不将其拉入默认回忆。感谢 @davemorin。 - 内存/做梦：在 Dreams 中添加 Dream Diary 表面，简化面向用户的做梦配置为 `enabled` 加可选 `frequency`，将阶段视为文档/UI 中的实现细节，并保持龙虾动画可见于日记内容上方。感谢 @vignesh07。 - 提示缓存：跨传输回退保持提示前缀更可重用，实现确定性 MCP 工具排序、压缩、嵌入图像历史、规范化系统提示指纹、`openclaw status --verbose` 缓存诊断，并从智能体系统提示中移除重复的带内工具清单，以便后续轮次更可靠地命中缓存。（#58036, #58037, #58038, #59054, #60603, #60691）感谢 @bcherny 和 @vincentkoc。 - 智能体/缓存诊断：添加提示缓存中断诊断，通过嵌入式运行程序路径跟踪实时缓存场景，并在 `openclaw status --verbose` 中显式显示缓存重用情况。感谢 @vincentkoc。 - 智能体/缓存：通过规范化等效的结构化提示空白、行尾、钩子添加的系统上下文以及运行时能力排序，稳定缓存相关系统提示指纹，以便语义未变的提示更可靠地复用 KV/缓存。感谢 @vincentkoc。 - 智能体/工具提示：从智能体系统提示中移除重复的带内工具清单，使工具调用模型将结构化工具定义作为唯一事实来源，从而提高提示稳定性并减少过时的工具指导。 - 配置/模式：为导出的 `openclaw config schema` JSON 模式添加字段标题和描述，以便编辑器、智能体和其他模式消费者接收相同的配置帮助元数据。（#60067）感谢 @solavrc。 - 提供商/CLI：移除捆绑的 CLI 文本提供商后端和 `agents.defaults.cliBackends` 表面，同时保留 ACP harness 会话和 Gemini 媒体理解的原生捆绑提供商。 - Matrix/执行审批：澄清不可用审批回复，使 Matrix 不再声称聊天审批不受支持，当原生执行审批仅未配置时。（#61424）感谢 @gumadeiras。 - 文档/IRC：将公共 IRC 主机名示例替换为 `irc.example.com`，并建议使用私有服务器进行机器人协调，同时列出常见公共网络供 intentional 使用。 - 内存/做梦：在将附近每日笔记行分组为短连贯块后再进行做梦暂存，以便来自近期笔记的一次性上下文以更好的证据和更少的行级噪声到达 REM/深睡阶段。 - 内存/做梦：从每日笔记块前缀中删除通用日期/日标题，同时保留有意义的节标签，以便暂存的片段更清晰、更可重用。（#61597）感谢 @mbelinky。 - 插件/Lobster：在进程中运行捆绑的 Lobster 工作流，而非生成外部 CLI，减少传输开销并解锁原生运行时集成。（#61523）感谢 @mbelinky。 - 插件/Lobster：强化托管恢复验证，使无效的 TaskFlow 恢复调用更早失败，并缓存每个运行程序的嵌入式运行时加载，同时保持失败加载可重试。（#61566）感谢 @mbelinky。 ### 修复 - 安全：保持仅限插件的工具允许列表的严格限制；`/allowlist add` 和 `/allowlist remove` 需要所有者访问权限；`before_tool_call` 钩子崩溃时失败即拒绝；更早阻止浏览器 SSRF 重定向绕过；并将非交互式认证选择推断范围限制在已捆绑且已信任的插件内。（#58476, #59836, #59822, #58771, #59120）感谢 @eleqtrizit 和 @pgondhi987。 - 提供商/OpenAI：使 GPT-5 和 Codex 运行更早采取行动，默认降低详细度，工具工作时显示可见进度，并在一轮仅叙述计划而非采取行动时进行一次性重试。 - 提供商/OpenAI 和回复传递：在支持的地方保留原生 `reasoning.effort: "none"` 和严格模式，跨重放和 Gateway `/v1/responses` 层添加 GPT-5.4 助手 `phase` 元数据，并将评论缓冲直到 `final_answer`，以便 Web 聊天、会话预览、嵌入式回复和 Telegram 部分停止泄露规划文本。修复 #59150, #59643, #61282。 - Telegram：修复模型选择器中的当前模型检查，HTML 格式化非默认 `/model` 确认，显式主题回复，跨重启的持久反应所有权，下载失败时的标题媒体占位符和 `file_id` 保留，以及升级安装的入站图像读取。（#60384, #60042, #59634, #59207, #59948, #59971）感谢 @sfuminya、@GitZhangChi、@dashhuang、@samzong、@v1p0r 和 @neeravmakwana。 - Telegram：恢复 DM 语音笔记预飞行转录，以便直接消息音频停止作为原始 `<media:audio>` 占位符到达。（#61008）感谢 @manueltarouca。 - Telegram/推理：仅当会话明确为 `reasoning:stream` 时才创建 Telegram 推理预览通道，以便流式回复中隐藏的 `停止在普通会话上显示为聊天预览。感谢 @vincentkoc。 - Telegram/原生命令菜单：在丢弃命令前修剪长菜单描述，以便亚 100 命令集仍能适应 Telegram 的载荷预算，并保持更多 `/` 条目可见。（#61129）感谢 @neeravmakwana。 - Discord：保持 REST、webhook 和监控流量在配置的代理上，保留仅组件的媒体发送，尊重 `@everyone` 和 `@here` 提及门控，在活动账户上保持 ACK 反应，并拆分语音连接/播放超时，以便自动加入更可靠。（#57465, #60361, #60345）感谢 @geekhuashan。 - Discord/回复标签：从预览文本中剥离泄露的 `[[reply_to_current]]` 控制标签，并在最终传递期间尊重显式回复标签线程，以便 Discord 回复保持附加到触发消息，而不是将回复元数据打印到聊天中。 - Discord/回复：用 `replyToMode: "batched"` 替换未发布的 `replyToOnlyWhenBatched` 标志，以便原生回复引用仅在去抖动多消息轮次上附加，而显式回复标签仍有效。 - Discord/图像生成：在工具输出中包含真实的生成 `MEDIA:` 路径，避免重复的纯输出媒体重新排队，并在最终回复传递前将易失的工作区生成的媒体持久化到持久出站媒体，以便生成的图像回复停止指向缺失的本地文件。 - Slack：将实时 DM 回复路由回具体的入站 DM 频道，同时保持持久路由元数据用户范围，以便正常助手回复停止在配对和系统消息仍到达时消失。（#59030）感谢 @afurm。 - WhatsApp：恢复 `channels.whatsapp.blockStreaming` 并在重新连接后重置看门狗超时，以便安静聊天停止陷入重新连接循环。（#60007, #60069）感谢 @MonkeyLeeT 和 @mcaxtr。 - Android/对话模式：当语音被显式停止时取消飞行中的 `talk.speak` 播放，并通过保持回复路由和嵌入式传输覆盖与当前播放路径对齐，在节点范围和网关支持的会话上恢复语音回复。（#60306, #61164, #61214） - 语音通话/OpenAI：将完整插件配置传递到实时转录提供商解析，以便流式通话能够再次发现捆绑的 OpenAI 实时转录提供商。修复 #60936。感谢 @sliekens 和 @vincentkoc。 - Matrix/执行审批：将播种的审批反应锚定到主 Matrix 提示事件，从事件元数据而非提示文本解析它们，并正确清理分块审批提示。（#60931）感谢 @gumadeiras。 - Matrix：通过重新创建秘密存储更可靠地恢复，当秘密存储或恢复密钥缺失时，在修复和备份重置期间，在持久化期间保持加密快照锁，并显示显式的过大附件标记。（#59846, #59851, #60599, #60289）感谢 @al3mart、@emonty 和 @efe-arv。 - Matrix/DM 会话：添加 `channels.matrix.dm.sessionScope`、共享会话冲突通知和对齐的出站会话重用，以便单独的 Matrix DM 房间在配置时能保持不同的上下文。（#61373）感谢 @gumadeiras。 - Matrix：将遗留顶层 `avatarUrl` 移动到多账户提升期间的默认账户，并保持环境支持的账户设置头像配置持久化。（#61437）感谢 @gumadeiras。 - MS Teams：通过 Graph API 下载内联 DM 图像，并在主动回退中保留频道回复线程。（#52212, #55198）感谢 @Ted-developer 和 @hyojin。 - MS Teams：用 `httpServerAdapter` 替换已弃用的 Teams SDK HttpPlugin 存根，以便重复的网关弃用警告停止触发，且 Express 5 兼容性解决方法保留在受支持的 SDK 路径上。（#60939）感谢 @coolramukaka-sys。 - 控制 UI/聊天：在聊天头部和移动聊天设置中添加每会话思维级别选择器，并保持浏览器捆绑包在 UI 本地思维/会话密钥助手上，以便 Safari 不再在渲染聊天控件前因仅 Node 导入而崩溃。 - 控制 UI：在仅工具执行期间保持停止可见，保留待发送忙碌状态，并在查询更改时立即清除过时的 ClawHub 搜索结果。（#54528, #59800, #60267）感谢 @chziyue 和 @frankekn。 - 控制 UI/头像：在提供 `/avatar/:agentId` 时尊重 `ui.assistant.avatar`，以便外观 UI 头像路径停止回退到首字母占位符。（#60778）感谢 @hannasdev。 - 控制 UI/cron：在刷新进行中时高亮 Cron 刷新按钮，以便即使先前数据保留在屏幕上，页面的加载状态也保持可见。（#60394）感谢 @coder-zhuzm。 - 控制 UI/概览：防止网关访问令牌/密码可见性切换按钮在窄宽度时重叠其输入。（#56924）感谢 @bbddbb1。 - 自动回复：跨预飞行压缩、会话轮换、CLI 支持运行和网关重启处理统一回复生命周期所有权，以便 `/stop` 和同会话重叠检查针对正确的活动轮次，且重启中断的轮次返回重启通知而非被静默丢弃。（#61267）感谢 @dutifulbob。 - 回复传递：防止在 `text_end` 频道上重复块回复，以便发出显式文本结束边界的提供商不再双送相同的最终消息。（#61530） - 网关/启动：当未设置时默认 `gateway.mode` 为 `local`，在 Windows 和 macOS 上检测网关锁文件中的 PID 回收，并显示启动进度，以便健康重启停止因陈旧锁而被阻塞。（#54801, #60085, #59843）感谢 @BradGroux 和 @TonyDerek-dot。 - 网关/macOS：让 launchd `KeepAlive` 再次拥有进程内网关重启，添加短的受控退出延迟，以便快速重启避免 launchd 崩溃循环卸载，同时 `openclaw gateway restart` 仍同步报告真实的 LaunchAgent 错误。 - 网关/macOS：如果 `launchctl kickstart -k` 在重启期间卸载它，则重新引导 LaunchAgent，以便失败重启不会将网关留空直到手动修复。 - 网关/macOS：在 `openclaw gateway start` 和 `restart` 期间恢复已安装但未加载的 LaunchAgents，同时在重启恢复期间仍优先考虑实时非托管网关。（#43766）感谢 @HenryC-3。 - 网关/Windows 计划任务：在重新安装时保留任务计划程序设置，当 `/Run` 未启动时大声失败，并准确报告快速失败重启而非假装它们在 60 秒后超时。（#59335）感谢 @tmimmanuel。 - Windows/重启：当计划任务从未注册时，回退到已安装的启动条目启动器，以便 `/restart` 能够在 Windows 设置上重新启动网关，其中 `schtasks` 安装在新手引导期间回退。（#58943）感谢 @imechZhangLY。 - Windows/重启：在 Windows 自重启前清理过时的网关监听器，并将监听器和 argv 探测失败视为不确定，以便计划任务重新启动不再陷入 `EADDRINUSE` 重试循环。（#60480）感谢 @arifahmedjoy。 - 更新/npm：偏好拥有已安装全局 OpenClaw 前缀的 npm 二进制文件，以便混合 Homebrew-plus-nvm 设置更新正确的安装。（#60153）感谢 @jayeshp19。 - 智能体/音乐和视频生成：添加 `tools.media.asyncCompletion.directSend` 作为已完成异步媒体任务的直接传递路径，同时保持传统的请求者会话唤醒/模型传递流为默认。 - CLI/技能 JSON：将 `skills list --json`、`skills info --json` 和 `skills check --json` 输出路由到 stdout 而非 stderr，以便机器可读消费者在预期流上再次接收 JSON。（#60914；修复 #57599；来自贡献者 PR #57611，由 @Aftabbs 提交）感谢 @Aftabbs。 - CLI/Commander：为参数和帮助错误路径保留 Commander 计算的退出代码，并在回归测试中覆盖用户 argv 解析模式，以便无效 CLI 调用在退出被拦截时不再报告成功。（#60923）感谢 @Linux2010。 - Cron：在第一次网关重启时重放中断的重复作业，而非等待第二次重启。（#60583）感谢 @joelnishanth。 - Cron：当未配置显式 `failureDestination` 时，通过作业的主要传递通道和与成功传递相同的会话上下文发送失败通知。（#60622）感谢 @artwalker。 - 执行/远程技能：当当前执行策略无法路由到节点时，停止广告 `exec host=node`，并澄清被阻止的执行主机覆盖错误，同时显示请求的主机和允许的配置路径。 - 智能体/Claude CLI/安全：清除继承的 Claude Code 配置根和插件根环境覆盖，如 `CLAUDE_CONFIG_DIR` 和 `CLAUDE_CODE_PLUGIN_*`，以便 OpenClaw 启动的 Claude CLI 运行无法被静默指向具有不同钩子、插件或认证上下文的不同 Claude 配置/插件树。感谢 @vincentkoc。 - 智能体/Claude CLI/安全：清除继承的 Claude Code 提供商路由和托管认证环境覆盖，并将 OpenClaw 启动的 Claude CLI 运行标记为托管主机，以便 Claude CLI 后门会话无法被静默重定向到代理、Bedrock、Vertex、Foundry 或父托管令牌上下文。感谢 @vincentkoc。 - 智能体/Claude CLI/安全：即使在自定义后端参数覆盖下，也强制托管主机 Claude CLI 后门运行到 `--setting-sources user`，以便仓库本地 `.claude` 项目/本地设置、钩子和插件发现不会在非交互式 OpenClaw 会话中静默执行。感谢 @vincentkoc。 - 智能体/Claude CLI：将格式错误的裸 `--permission-mode` 后端覆盖视为缺失并安全回退到 `bypassPermissions`，以便自定义 `cliBackends.claude-cli.args` 安全配置不能意外消耗下一个标志作为虚假权限模式。感谢 @vincentkoc。 - 网关/设备配对：要求非管理员配对设备会话仅管理自己的设备以进行令牌轮换/撤销和配对设备移除，阻止配对范围内会话内的跨设备令牌盗窃。（#50627）感谢 @coygeek。 - 网关/插件路由：在仅写入回退范围上保持网关认证插件运行时路由，除非受信任的代理调用者显式声明更窄的 `x-openclaw-scopes`，以便插件 HTTP 处理程序不再在缺失或不可信 HTTP 范围标头上铸造管理员级运行时范围。（#59815）感谢 @pgondhi987。 - 构建/类型：修复 Node `createRequire(...)` 助手类型，以便提供商运行时惰性加载再次干净编译，且 `pnpm build` 不再在 Pi 嵌入式提供商错误模式路径中失败。 - 网关/安全：按规范化来源范围环回浏览器来源认证节流，以便一个本地主机控制 UI 选项卡在重复认证失败后不能锁定不同的本地主机浏览器来源。 - 网关/认证：序列化每个客户端的异步共享秘密认证尝试，以便并发 Tailscale 能力失败不能超出预期认证速率限制预算。感谢 @Telecaster2147。 - 设备配对/安全：将非操作员设备范围检查绑定到请求的角色前缀，以便引导验证不能通过 `node` 认证赎回 `operator.*` 范围。（#57258）感谢 @jlapenna。 - 设备配对：拒绝将设备令牌轮换到配对期间从未批准的角色，并将重新连接角色检查绑定到配对设备的批准角色集。（#60462）感谢 @eleqtrizit。 - 网关/设备认证：仅对缓存令牌重新连接重用缓存的设备令牌范围，同时保持显式 `deviceToken` 范围请求和空缓存回退完整，以便重新连接保留 `operator.read` 而不破坏显式认证流。（#46032）感谢 @caicongyang。 - 移动配对/安全：当网关配对范围缺失时，对内部 `/pair` 设置代码颁发、清理和批准路径失败关闭，并在内部命令路径上保持批准时请求范围强制。（#55996）感谢 @coygeek。 - 移动配对/引导：将 QR 引导交接令牌绑定到移动安全合约，以便节点交接保持无范围，操作员交接丢弃混合的 `node.*`、`operator.admin` 和 `operator.pairing` 范围。 - 移动配对/Android：收紧安全端点处理，以便 Tailscale 和公共远程设置拒绝明文端点，私有 LAN 配对仍工作，合并角色批准铸造节点和操作员设备令牌，且引导令牌在节点自动配对后持续到操作员批准完成。（#60128, #60208, #60221）感谢 @obviyus。 - Android/画布安全：在转发画布桥操作前要求精确的规范化 A2UI URL 匹配，拒绝查询不匹配和后代路径，同时允许仅片段的 A2UI 导航。 - 群晖聊天/安全：默认将低级别 HTTPS 助手 TLS 验证设为开启，以便助手/API 默认与 shipped 安全账户默认匹配，且仅显式 `allowInsecureSsl: true` 选择退出。 - 群晖聊天/安全：通过共享恒定时间秘密助手路由 webhook 令牌比较，以便与其他捆绑插件一致。 - 插件/市场：阻止远程市场符号链接逃逸，而不破坏普通本地市场安装路径。（#60556）感谢 @eleqtrizit。 - Telegram/本地 Bot API：尊重 `channels.telegram.apiRoot` 进行缓冲媒体下载，添加 `channels.telegram.network.dangerouslyAllowPrivateNetwork` 用于受信任的假 IP 设置，并在读取绝对 Bot API `file_path` 值前要求 `channels.telegram.trustedLocalFileRoots`。（#59544, #60705）感谢 @SARAMALI15792 和 @obviyus。 - 出站/消毒剂：从共享的用户可见助手文本中剥离泄露的 `<tool_call>`、`<function_calls>` 和模型特殊令牌，包括截断的工具调用流，以便内部脚手架不再跨表面渗入回复。（#60619）感谢 @oliviareid-svg。 - 智能体/错误：当本地会话或转录写入因 `ENOSPC`/磁盘满失败时，显示显式磁盘满消息，以便这些运行停止退化为 opaque `NO_REPLY` 风格失败。感谢 @vincentkoc。 - 执行审批：从主机执行中移除启发式命令混淆门控，以便网关和节点运行依赖显式策略、允许列表和严格内联评估规则。 - 智能体/工具结果：将实时工具结果持久化和溢出恢复截断上限设为 40k 字符，以便过大的工具输出保持有界而不完全丢弃最近上下文。 - Discord/视频回复：将文本加视频传递拆分为文本回复后跟仅媒体发送，并让实时提供商认证检查尊重清单声明的 API 密钥环境变量，如 `MODELSTUDIO_API_KEY`。 - 配置/所有设置：当敏感字段为空时保持原始配置视图完整，而不是损坏或丢弃渲染快照。（#28214）感谢 @solodmd。 - 插件 SDK/外观：在插件 ID 跟踪重新进入配置加载前回填捆绑插件外观哨兵，以便 CLI/提供商启动不再在捆绑插件重新进入期间出现 `shouldNormalizeGoogleProviderConfig is not a function` 或其他空外观读取错误。感谢 @adam91holt。 - 插件/外观：在跟踪插件解析重新进入配置加载前回填外观哨兵，以便外观导出在循环提供商规范化期间保持已定义。（#61180）感谢 @adam91holt。 - QA 实验室：恢复类型模拟 OpenAI 网关配置连接，以便 QA 实验室配置助手再次干净编译，且 `pnpm check` / `pnpm build` 保持绿色。 - Discord/图像生成：在工具输出中包含真实的生成 `MEDIA:` 路径并避免重复纯输出媒体重新排队，以便 Discord 图像回复停止指向缺失的本地文件。 - Slack：将实时 DM 回复路由回具体的入站 DM 频道，同时保持持久路由元数据用户范围，以便正常助手回复停止在配对和系统消息仍到达时消失。（#59030）感谢 @afurm。 - Discord/回复标签：从预览文本中剥离泄露的 `[[reply_to_current]]` 控制标签，并在最终传递期间尊重显式回复标签线程，以便 Discord 回复保持附加到触发消息，而不是将回复元数据打印到聊天中。 - Telegram：修复模型选择器中的当前模型检查，HTML 格式化非默认 `/model` 确认，显式主题回复，跨重启的持久反应所有权，下载失败时的标题媒体占位符和 `file_id` 保留，以及升级安装的入站图像读取。（#60384, #60042, #59634, #59207, #59948, #59971）感谢 @sfuminya、@GitZhangChi、@dashhuang、@samzong、@v1p0r 和 @neeravmakwana。 - Telegram：恢复 DM 语音笔记预飞行转录，以便直接消息音频停止作为原始 `<media:audio>` 占位符到达。（#61008）感谢 @manueltarouca。 - Telegram/推理：仅当会话明确为 `reasoning:stream` 时才创建 Telegram 推理预览通道，以便流式回复中隐藏的 `<think>` 跟踪停止在普通会话上显示为聊天预览。感谢 @vincentkoc。 - Telegram/原生命令菜单：在丢弃命令前修剪长菜单描述，以便亚 100 命令集仍能适应 Telegram 的载荷预算，并保持更多 `/` 条目可见。（#61129）感谢 @neeravmakwana。 - 飞书/推理：仅当会话明确为 `reasoning:stream` 时才公开流式推理预览，以便隐藏推理跟踪不在普通流式会话上显示。感谢 @vincentkoc。 - Discord：保持 REST、webhook 和监控流量在配置的代理上，保留仅组件的媒体发送，尊重 `@everyone` 和 `@here` 提及门控，在活动账户上保持 ACK 反应，并拆分语音连接/播放超时，以便自动加入更可靠。（#57465, #60361, #60345）感谢 @geekhuashan。 - WhatsApp：恢复 `channels.whatsapp.blockStreaming` 并在重新连接后重置看门狗超时，以便安静聊天停止陷入重新连接循环。（#60007, #60069）感谢 @MonkeyLeeT 和 @mcaxtr。 - 内存：保持 `memory-core` 内置嵌入注册在已注册路径上，以便选择 `memory-core` 不再递归通过插件发现并在启动期间崩溃。（#61402）感谢 @ngutman。 - 智能体/工具结果：保持大型 `read` 输出可见更久，当较旧工具输出可吸收溢出预算时保留最新 `read` 输出，并在用压缩存根替换新鲜 `read` 前回退到 Pi 的正常溢出压缩/重试路径。感谢 @vincentkoc。 - 内存/QMD：偏好现代 `qmd collection add --glob`，接受较新的单行 JSON 命中元数据同时保持遗留行字段，刷新 QMD 文档/doctor 安装指南和模型覆盖指南，并保持较旧 QMD 版本工作。感谢 @vincentkoc。 - MS Teams：通过 Graph API 下载内联 DM 图像，并在主动回退中保留频道回复线程。（#52212, #55198）感谢 @Ted-developer 和 @hyojin。 - MS Teams：用 `httpServerAdapter` 替换已弃用的 Teams SDK HttpPlugin 存根，以便重复的网关弃用警告停止触发，且 Express 5 兼容性解决方法保留在受支持的 SDK 路径上。（#60939）感谢 @coolramukaka-sys。 - Matrix/执行审批：将播种的审批反应锚定到主 Matrix 提示事件，从事件元数据而非提示文本解析它们，并正确清理分块审批提示。（#60931）感谢 @gumadeiras。 - Matrix：通过重新创建秘密存储更可靠地恢复，当秘密存储或恢复密钥缺失时，在修复和备份重置期间，在持久化期间保持加密快照锁，并显示显式的过大附件标记。（#59846, #59851, #60599, #60289）感谢 @al3mart、@emonty 和 @efe-arv。 - Android/对话模式：当语音被显式停止时取消飞行中的 `talk.speak` 播放，以便过时回复停止在打断或手动停止后开始。（#61164）感谢 @obviyus。 - Android/对话模式：通过保持回复路由同步到解析的节点会话密钥并在回复播放期间暂停麦克风捕获，在节点范围会话上恢复语音助手回复。（#60306）感谢 @MKV21。 - Android/对话模式：通过更新嵌入式运行程序传输覆盖到当前智能体传输 API，在网关支持的对话模式会话上恢复语音回复。（#61214）感谢 @obviyus。 - 语音通话/OpenAI：将完整插件配置传递到实时转录提供商解析，以便流式通话能够再次发现捆绑的 OpenAI 实时转录提供商。修复 #60936。感谢 @sliekens 和 @vincentkoc。 - 控制 UI/聊天：在聊天头部和移动聊天设置中添加每会话思维级别选择器，并保持浏览器捆绑包在 UI 本地思维/会话密钥助手上，以便 Safari 不再在渲染聊天控件前因仅 Node 导入而崩溃。 - 控制 UI：在仅工具执行期间保持停止可见，保留待发送忙碌状态，并在查询更改时立即清除过时的 ClawHub 搜索结果。（#54528, #59800, #60267）感谢 @chziyue 和 @frankekn。 - 控制 UI/头像：在提供 `/avatar/:agentId` 时尊重 `ui.assistant.avatar`，以便外观 UI 头像路径停止回退到首字母占位符。（#60778）感谢 @hannasdev。 - 控制 UI/cron：在刷新进行中时高亮 Cron 刷新按钮，以便即使先前数据保留在屏幕上，页面的加载状态也保持可见。（#60394）感谢 @coder-zhuzm。 - 控制 UI/概览：防止网关访问令牌/密码可见性切换按钮在窄宽度时重叠其输入。（#56924）感谢 @bbddbb1。 - CLI/技能 JSON：将 `skills list --json`、`skills info --json` 和 `skills check --json` 输出路由到 stdout 而非 stderr，以便机器可读消费者在预期流上再次接收 JSON。（#60914；修复 #57599；来自贡献者 PR #57611，由 @Aftabbs 提交）感谢 @Aftabbs。 - CLI/Commander：为参数和帮助错误路径保留 Commander 计算的退出代码，并在回归测试中覆盖用户 argv 解析模式，以便无效 CLI 调用在退出被拦截时不再报告成功。（#60923）感谢 @Linux2010。 - Cron：在第一次网关重启时重放中断的重复作业，而非等待第二次重启。（#60583）感谢 @joelnishanth。 - Cron：当未配置显式 `failureDestination` 时，通过作业的主要传递通道和与成功传递相同的会话上下文发送失败通知。（#60622）感谢 @artwalker。 - 实时模型切换：仅将显式用户驱动的模型更改视为待处理的实时切换，以便回退轮换、心跳覆盖和压缩不再在发出 API 调用前触发 `LiveSessionModelSwitchError`。（#60266）感谢 @kiranvk-2011。 - 执行审批：在允许列表模式中复用持久的精确命令 `allow-always` 审批，以便相同重运行停止提示，并收紧 Windows 解释器/路径审批处理，以便包装器和畸形路径情况更一致地失败关闭。（#59880, #59780, #58040, #59182）感谢 @luoyanglang、@SnowSky1 和 @pgondhi987。 - 节点执行审批：将节点主机 `system.run` 审批绑定到跨异步转发的准备执行计划，以便可变脚本操作数仍获得审批时绑定和漂移重新验证，而不是回退到无绑定执行。 - 智能体/执行审批：让 `exec-approvals.json` 智能体安全覆盖更严格的网关工具默认值，以便批准的副智能体可以使用 `security: “full”` 而不再回退到允许列表执行。（#60310）感谢 @lml2468。 - 智能体/执行：恢复 `host=node` 路由用于节点固定和 `host=auto` 会话，同时仍阻止沙盒化 `auto` 会话跳转到网关。（#60788）感谢 @openperf。 - 执行/心跳：使用规范的 `exec-event` 唤醒原因作为 `notifyOnExit`，以便后台执行完成仍在 `HEARTBEAT.md` 为空或仅注释时触发后续轮次。（#41479）感谢 @rstar327。 - 心跳：当目标会话通道已忙碌时跳过唤醒传递，以便待处理事件被重试而非过早耗尽。（#40526）感谢 @lucky7323。 - 群聊/智能体提示：告诉模型最小化空行并使用正常聊天风格间距，以便群组回复避免文档风格空行格式。 - 提供商/OpenAI GPT：将短审批轮次（如 `ok do it` 和 `go ahead`）视为即时操作轮次，并修剪过度类似备忘录的 GPT-5 聊天确认，以便 OpenAI 回复默认更短、更会话化。 - 提供商/OpenAI Codex：将原生 `contextWindow` 与运行时 `contextTokens` 分离，将默认有效上限保持在 `272000`，并在 `models.providers.*.models[]` 上公开每模型 `contextTokens` 覆盖。 - 提供商/OpenAI 兼容 WS：当提供商省略或为零 `total_tokens` 时，从规范化使用情况计算回退令牌总数，以便 DashScope 兼容会话在别名规范化后停止存储零总数。（#54940）感谢 @lyfuci。 - 智能体/OpenAI：将 Claude 兼容文件工具模式标记为 `additionalProperties: false`，以便直接 OpenAI GPT-5 路由停止因无效严格模式错误而拒绝 `read` 工具。 - 智能体/OpenAI：当工具模式不严格兼容时回退到 `strict: false`，并将空对象工具模式规范化为包含 `required: []`，以便直接 GPT-5 路由停止因无效严格模式错误（如 `read` 中缺少 `path`）而失败。 - 智能体/GPT：为嵌入式运行添加显式工作项生命周期事件，使用它们更可靠地显示真实进度，并停止将工具启动轮次计为仅规划重试。 - 插件/OpenAI：通过 `/images/edits` 多部分上传启用 `gpt-image-1` 参考图像编辑，并在未提供显式 `size` 或 `resolution` 时停止推断不支持的解析覆盖。 - 智能体/重放：从重放历史消毒中移除格式错误的助手内容规范化修复，而非将该遗留修复路径扩展到重放验证。 - 插件/OpenAI：调整 OpenAI 提示覆盖以实现实时聊天节奏，以便 GPT 回复默认更短、更人性化、更少大段文本。 - 提供商/兼容：停止在代理和自定义 OpenAI 兼容路由上强制仅 OpenAI 默认，跨 Anthropic 兼容、Moonshot、Mistral、ModelStudio、OpenRouter、xAI 和 Z.ai 端点保留原生供应商特定推理/工具/流行为，并将 GitHub Copilot Claude 模型通过 Anthropic Messages 而非 OpenAI Responses 路由。 - 提供商/GitHub Copilot：在运行时模型请求和 GitHub 令牌交换上发送 IDE 身份标头，以便 IDE 认证的 Copilot 运行停止因缺少 `Editor-Version` 而失败。（#60641）感谢 @VACInc 和 @vincentkoc。 - 提供商/OpenRouter 故障转移：将 `403 “Key limit exceeded”` 消费限制响应分类为计费，以便模型故障转移继续而非在通用认证上停止。（#59892）感谢 @rockcent。 - 提供商/Anthropic：在实时 Claude CLI 凭证上保持 `claude-cli/*` 认证，避免持久化过时的持有者令牌配置文件，并在非交互式 Claude CLI 设置期间抑制 macOS 钥匙串提示。（#61234）感谢 @darkamenosa。 - 提供商/Anthropic：当 Claude CLI 认证成为默认时，写入真实的 `claude-cli` 认证配置文件，以便本地和网关智能体运行能立即使用 Claude CLI 而不会出现缺失 API 密钥失败。感谢 @vincentkoc。 - 提供商/Anthropic Vertex：在 Vertex AI 端点上尊重 `cacheRetention: “long”` 配合真实的 1 小时提示缓存 TTL，并像直接 Anthropic 一样默认 `anthropic-vertex` 缓存保留。（#60888）感谢 @affsantos。 - 智能体/Anthropic：在直接 Anthropic 和 Anthropic Vertex 路径上保留原生 `toolu_*` 重放 ID，以便缓存敏感历史停止重写已知有效的 Anthropic 工具使用 ID。（#52612） - 提供商/Google：通过自动创建、重用和刷新 `cachedContents`，为直接 Gemini 系统提示添加模型级 `cacheRetention` 支持，以便 Google AI Studio 运行支持提示缓存。（#51372）感谢 @rafaelmariano-glitch。 - Google Gemini CLI 认证：通过扫描捆绑包文件中的 Gemini OAuth 客户端配置来检测捆绑的 npm 安装，以便 `npm install -g @google/gemini-cli` 布局再次工作。（#60486）感谢 @wzfmini01。 - Google Gemini CLI 认证：从本地 Gemini 设置检测个人 OAuth 模式并跳过这些登录的 Code Assist 项目发现，以便个人 Google 账户停止因 `loadCodeAssist 400 Bad Request` 而失败。（#49226）感谢 @bobworrall。 - Google Gemini CLI 认证：改进 Windows nvm 和 Homebrew libexec 安装的 OAuth 凭证发现，并对齐 Code Assist 元数据，以便 Gemini 登录在打包 CLI 布局上停止失败。（#40729）感谢 @hughcube。 - Google Gemini CLI 模型：通过让捆绑 CLI 提供商从 Google 模板克隆它们，为稳定的 `gemini-2.5-*` 模型 ID 添加前向兼容支持，以便 `gemini-2.5-flash-lite` 和相关配置模型停止显示为缺失。（#35274）感谢 @mySebbe。 - Google 图像生成：为 Gemini 图像请求禁用固定 DNS，并尊重共享提供商 HTTP 助手中的显式 `pinDns` 覆盖，以便代理支持的图像生成再次工作。（#59873）感谢 @luoyanglang。 - 提供商/Microsoft Foundry：在规范化的 Foundry 部署上保留显式图像能力，修复网关和运行时路径上过时的 GPT/o 系列仅文本模型元数据，并防止未知回退模型借用不相关的图像支持。 - 提供商/Model Studio：即使配置在通用提供商密钥下，也保留 DashScope 兼容端点的原生流式使用报告，以便流式令牌总数停止粘在零。（#52395）感谢 @IVY-AI-gif。 - 提供商/Z.AI：保留显式注册的 `glm-5-*` 变体（如 `glm-5-turbo`），而非用通用 GLM-5 前向兼容填充拦截它们。（#48185）感谢 @haoyu-haoyu。 - Amazon Bedrock/aws-sdk 认证：当不存在 AWS 认证环境变量时，停止注入虚假的 `AWS_PROFILE` apiKey 标记，以便实例角色和其他默认链设置保持工作而不污染提供商配置。（#61194）感谢 @wirjo。 - 智能体/Kimi 工具调用修复：当后续畸形增量失败重新评估时，保留流式工具调用上已存在的工具参数，同时仍在 `toolcall_end` 前丢弃仅修复的状态。 - 插件/Kimi Coding：解析标记的工具调用并保留 Anthropic 原生工具负载，以便 Kimi 编码端点执行工具而非回显原始标记。（#60051, #60391）感谢 @obviyus 和 @Eric-Guo。 - 媒体理解：自动为视觉路由注册图像能力提供商，以便具有图像模型的自定义 GLM 样式提供商 ID 停止因“未注册媒体理解提供商”而失败。（#51418）感谢 @xydt-610。 - 插件/媒体理解：默认启用捆绑的 Groq 和 Deepgram 提供商，以便配置的转录模型无需额外插件激活配置即可工作。（#59982）感谢 @yxjsxy。 - MiniMax/定价：在提供商目录中保持捆绑 MiniMax 高速定价 distinct，并在引导较旧 MiniMax 模型时保留较低的 M2.5 缓存读取定价。（#54214）感谢 @octo-patch。 - MiniMax：在捆绑的 `MiniMax-M2.7` 和 `MiniMax-M2.7-highspeed` 模型定义上广告图像输入，以便图像能力流能正确通过 M2.7 系列路由。（#54843）感谢 @MerlinMiao88888888。 - 模型/MiniMax：为隐式捆绑 MiniMax 提供商目录尊重 `MINIMAX_API_HOST`，以便中国托管的 API 密钥设置选择 `api.minimaxi.com/anthropic` 而无需手动提供商配置。（#34524）感谢 @caiqinghua。 - 使用情况/MiniMax：当 MiniMax 仅报告剩余百分比数据时，反转剩余风格 `usage_percent` 字段，以便使用情况条停止将几乎满的剩余配额显示为几乎耗尽的使用情况。 - 使用情况/MiniMax：让使用情况快照将 `minimax-portal` 和 MiniMax CN 别名视为相同的 MiniMax 配额表面，并优先存储的 MiniMax OAuth 而非回退到 Coding Plan 密钥。 - 使用情况/MiniMax：优先聊天模型 `model_remains` 条目并从 MiniMax 间隔时间戳推导 Coding Plan 窗口标签，以便 MiniMax 使用情况快照停止选择零预算媒体行和将 4h 窗口错误报告为 `5h`。（#52349）感谢 @IVY-AI-gif。 - 模型选择器/提供商：在设置期间将捆绑的 BytePlus 和 Volcengine 计划别名视为其原生提供商，并在认证前公开其捆绑标准/编码目录，以便设置能建议正确的模型。（#58819）感谢 @Luckymingxuan。 - 工具/web_search (Kimi)：当 `tools.web.search.kimi.baseUrl` 未设置时，继承原生 Moonshot 聊天 `baseUrl`（`.ai` / `.cn`），以便中国控制台密钥在同一主机上认证聊天。修复 #44851。（#56769）感谢 @tonga54。 - 智能体/Claude CLI：在自定义 `cliBackends.claude-cli.args` 覆盖默认值（包括运行时插件注册表激活前的回退解析）时保持非交互式 `--permission-mode bypassPermissions`，以便 cron 和心跳 Claude CLI 运行不回归到交互式批准模式。（#61114）感谢 @cathrynlavery 和 @thewilloftheshadow。 - 智能体/Claude CLI：将显式 `openclaw agent --session-id` 运行持久化在稳定会话密钥下，以便后续能复用存储的 CLI 绑定并恢复相同的底层 Claude 会话。 - 智能体/Claude CLI：持久化路由的 Claude 会话绑定，在 `/new` 和 `/reset` 上轮换它们，并保持实时 Claude CLI 模型切换在配置的 Claude 系列上移动，以便恢复的会话跟随真实活动线程和模型。感谢 @vincentkoc。 - 智能体/CLI 后端：当本地 CLI 登录状态或所选认证配置文件凭证更改时，使存储的 CLI 会话重用失效，以便重新登录和令牌轮换停止恢复过时会话。 - 智能体/Claude CLI/图像：复用稳定水合图像文件路径并保留共享媒体扩展（如 HEIC）在将图像引用传递给本地 CLI 运行时，以便 Claude CLI 图像提示停止 thrash KV 缓存前缀，且奇怪图像格式不回退到 `.bin`。感谢 @vincentkoc。 - 智能体/压缩：将助手工具调用和位移工具结果保持在相同压缩块中，以便严格摘要提供商停止拒绝孤立的工具对。（#58849）感谢 @openperf。 - 智能体/故障转移：按提供商范围 Anthropic `An unknown error occurred` 故障转移匹配，以便通用内部未知错误文本不再触发可重试超时回退。（#59325）感谢 @aaron-he-zhu。 - 智能体/副智能体：当副智能体在 `LiveSessionModelSwitchError` 后重试时，尊重允许列表验证、认证配置文件交接和会话覆盖状态。（#58178）感谢 @openperf。 - 智能体/运行时：使默认副智能体允许列表、继承的技能/工作区和重复会话 ID 解析行为更可预测，并在缺失参数工具错误中包含值形状提示。（#59944, #59992, #59858, #55317）感谢 @hclsys、@gumadeiras、@joelnishanth 和 @priyansh19。 - 智能体/配对：将完成宣布传递上下文与请求者会话回退合并，以便缺失 `to` 仍到达原始频道，并在节点角色设备配对批准中包含 `operator.talk.secrets` 到 CLI 默认操作员范围。（#56481）感谢 @maxpetrusenko。 - 智能体/调度：将后台现在工作转向自动完成唤醒，并将 `process` 轮询视为按需检查或干预而非默认完成处理。（#60877）感谢 @vincentkoc。 - 智能体/技能：将技能镜像到沙盒工作区时跳过 `.git` 和 `node_modules`，以便只读沙盒不复制仓库历史或依赖树。（#61090）感谢 @joelnishanth。 - ACP/智能体：为跨智能体 ACP 生成继承目标智能体工作区，并在继承的工作区不再存在时安全回退。（#58438）感谢 @zssggle-rgb。 - ACPX/Windows：在 Claude CLI 解析中保留反斜杠和绝对 `.exe` 路径，并快速失败于包装脚本目标，指导使用 `cmd.exe /c`、`powershell.exe -File` 或 `node <script>`。（#60689）感谢 @steipete。 - 认证/故障转移：在重试前持久化选择的覆盖，缩短 `auth_permanent` 锁定，且仅当真实认证更改时刷新 websocket/共享认证会话，以便重试和密钥轮换行为可预测。（#60404, #60323, #60387）感谢 @extrasmall0 和 @mappel-nv。 - 网关/频道：在后期插件注册混乱前固定初始启动频道注册，以便配置的频道保持可见，且 `channels.status` 停止在运行时插件加载后回退到空 `channelOrder` / `channels` 载荷。 - 提示缓存：在系统提示缓存边界前对稳定工作区项目上下文文件排序，并保持 `HEARTBEAT.md` 在系统提示缓存边界下，以便心跳混乱不使稳定项目上下文前缀失效。（#58979）感谢 @yozu 和 @vincentkoc。 - 提示缓存：将 Codex Responses 和 Anthropic Vertex 通过边界感知缓存整形路由，并在缓存跟踪中报告实际出站系统提示，以便缓存重用和未命中与提供商实际接收的对齐。感谢 @vincentkoc。 - 智能体/缓存：跨工具循环保留完整的 3 轮提示缓存图像窗口，使冲突的捆绑 MCP 工具定义确定，并在有效负载钩子替换后重新应用 Anthropic Vertex 缓存整形，以便 KV/缓存重用保持稳定。感谢 @vincentkoc。 - 状态/缓存：在转录回退中恢复 `cacheRead` 和 `cacheWrite`，以便 `/status` 在会话日志是唯一完整使用来源时保持显示缓存命中百分比。（#59247）感谢 @stuartsy。 - 状态/使用情况：让 `/status` 和 `session_status` 在会话元存储保持为零时回退到转录令牌总数，以便 LM Studio、Ollama、DashScope 和类似 OpenAI 兼容提供商停止显示 `Context: 0/...`。（#55041）感谢 @jjjojoj。 - Mattermost/配置模式：再次接受 `groups.*.requireMention`，以便现有 Mattermost 配置在升级后不再失败严格验证。（#58271）感谢 @MoerAI。 - Doctor/配置：通过深度结构相等性而非键顺序敏感的序列化比较规范化 `talk` 配置，以便 `openclaw doctor --fix` 停止重复报告/应用无操作 `talk.provider/providers` 规范化。（#59911）感谢 @ejames-dev。 - Anthropic CLI 引导：在非交互式 Claude CLI 设置期间也重写迁移的回退模型引用，以便引导和脚本化设置不再在将主模型切换到 `claude-cli/*` 后保留过时的 `anthropic/*` 回退。感谢 @vincentkoc。 - 模型/Anthropic CLI 认证：当 `openclaw models auth login --provider anthropic --method cli --set-default` 切换到 `claude-cli/*` 时替换迁移的 `agents.defaults.models` 允许列表，以便过时的 `anthropic/*` 条目不残留 beside 迁移的 Claude CLI 默认值。感谢 @vincentkoc。 - Doctor/Claude CLI：添加专用 Claude CLI 健康检查，以便 `openclaw doctor` 能在智能体运行失败前发现缺失的本地安装或损坏的认证。感谢 @vincentkoc。 - 插件/认证选择：在不递归保留替换的默认模型映射的情况下应用提供商拥有的认证配置补丁，以便 Anthropic Claude CLI 和类似迁移能在引导和设置期间有意交换模型允许列表，而非累积过时条目。感谢 @vincentkoc。 - 插件/引导：将点状插件 uiHint 路径（如 Brave `webSearch.mode`）写为嵌套插件配置，以便 `llm-context` 设置停止失败验证。（#61159）感谢 @obviyus。 - 插件/安装：在链接插件和钩子包探测间保留不安全覆盖标志，以便本地 `--link` 安装尊重文档覆盖行为。（#60624）感谢 @JerrettDavis。 - 插件/缓存：当调用者省略 `workspaceDir` 时，为提供商、web-search 和 web-fetch 快照加载继承活动网关工作区，以便兼容插件注册表和快照缓存不在网关拥有的运行时路径上缺失。（#61138）感谢 @jzakirov。 - 插件 SDK/上下文引擎：从 `openclaw/plugin-sdk` 导出缺失的上下文引擎结果和副智能体生命周期类型，以便上下文引擎插件能类型化 `ContextEngine` 实现而无需本地变通。（#61251）感谢 @DaevMithran。 - 任务/维护：将过时的 cron 和聊天支持的 CLI 任务行与实时 cron 作业和智能体运行所有权协调，而非将任何持久会话密钥视为任务仍在运行的证明。（#60310）感谢 @lml2468。 - 插件：在非激活快照和 CLI 元数据加载期间抑制信任警告噪音。（#61427）感谢 @gumadeiras。 - 智能体/视频生成：在严格配置验证和 `openclaw config set/get` 中接受 `agents.defaults.videoGenerationModel`，以便启用视频模型的网关使用 `video_generate` 后不再失败启动。 - Matrix/流式：为流式 Matrix 回复添加安静预览模式，保持遗留 `partial` 预览优先行为，并正确完成安静媒体标题，以便预览停止过早通知而不丢弃最终文本语义。（#61450）感谢 @gumadeiras。 - 网关/关闭：即使没有跟踪客户端也绑定 websocket 服务器关闭，以便网关重启停止挂起直到看门狗杀死进程。（#61565）感谢 @mbelinky。 - 控制 UI/多语言：本地化剩余的共享频道、实例、节点和网关确认字符串，以便仪表板停止混合翻译 UI 与硬编码英文标签。感谢 @vincentkoc。 - Discord/媒体：将默认入站和出站媒体上限提高到 `100MB`，以便 Discord 更接近 Telegram，更大的附件停止在旧低默认值上失败。 - Matrix：将直接传输请求保持在固定调度程序上，通过 undici 运行时获取路由，以便 Matrix 客户端在较新运行时恢复同步而不丢弃已验证的地址绑定。（#61595）感谢 @gumadeiras。 - 插件/外观：从注册表根解析全局安装的捆绑插件运行时外观，以便捆绑频道（如 LINE）在获胜插件安装位于全局扩展目录下且编码作用域文件夹名称时仍启动。（#61297）感谢 @openperf。

### Breaking - Config: remove legacy public config aliases such as `talk.voiceId` / `talk.apiKey`, `agents.*.sandbox.perSession`, `browser.ssrfPolicy.allowPrivateNetwork`, `hooks.internal.handlers`, and channel/group/room `allow` toggles in favor of the canonical public paths and `enabled`, while keeping load-time compatibility and `openclaw doctor --fix` migration support for existing configs. (#60726) Thanks @vincentkoc. ### Changes - Agents/video generation: add the built-in `video_generate` tool so agents can create videos through configured providers and return the generated media directly in the reply. - Agents/music generation: ignore unsupported optional hints such as `durationSeconds` with a warning instead of hard-failing requests on providers like Google Lyria. - Providers/ComfyUI: add a bundled `comfy` workflow media plugin for local ComfyUI and Comfy Cloud workflows, including shared `image_generate`, `video_generate`, and workflow-backed `music_generate` support, with prompt injection, optional reference-image upload, live tests, and output download. - Tools/music generation: add the built-in `music_generate` tool with bundled Google (Lyria) and MiniMax providers plus workflow-backed Comfy support, including async task tracking and follow-up delivery of finished audio. - Providers: add bundled Qwen, Fireworks AI, and StepFun providers, plus MiniMax TTS, Ollama Web Search, and MiniMax Search integrations for chat, speech, and search workflows. (#60032, #55921, #59318, #54648) - Providers/Amazon Bedrock: add bundled Mantle support plus inference-profile discovery and automatic request-region injection so Bedrock-hosted Claude, GPT-OSS, Qwen, Kimi, GLM, and similar routes work with less manual setup. (#61296, #61299) Thanks @wirjo. - Control UI/multilingual: add localized control UI support for Simplified Chinese, Traditional Chinese, Brazilian Portuguese, German, Spanish, Japanese, Korean, French, Turkish, Indonesian, Polish, and Ukrainian. Thanks @vincentkoc. - Plugins: add plugin-config TUI prompts to guided onboarding/setup flows, and add `openclaw plugins install --force` so existing plugin and hook-pack targets can be replaced without using the dangerous-code override flag. (#60590, #60544) - Control UI/skills: add ClawHub search, detail, and install flows directly in the Skills panel. (#60134) Thanks @samzong. - iOS/exec approvals: add generic APNs approval notifications that open an in-app exec approval modal, fetch command details only after authenticated operator reconnect, and clear stale notification state when the approval resolves. (#60239) Thanks @ngutman. - Matrix/exec approvals: add Matrix-native exec approval prompts with account-scoped approvers, channel-or-DM delivery, and room-thread aware resolution handling. (#58635) Thanks @gumadeiras. - Channels/context visibility: add configurable `contextVisibility` per channel (`all`, `allowlist`, `allowlist_quote`) so supplemental quote, thread, and fetched history context can be filtered by sender allowlists instead of always passing through as received. - Providers/request overrides: add shared model and media request transport overrides across OpenAI-, Anthropic-, Google-, and compatible provider paths, including headers, auth, proxy, and TLS controls. (#60200) - Providers/OpenAI: add forward-compat `openai-codex/gpt-5.4-mini`, an opt-in GPT personality, and provider-owned GPT-5 prompt contributions so Codex/GPT runs stay cache-stable and compatible with bundled catalog lag. - Agents/Claude CLI: expose OpenClaw tools to background Claude CLI runs through a loopback MCP bridge and switch bundled runs to stdin + `stream-json` partial-message streaming so prompts stop riding argv, long replies show live progress, and final session/usage metadata still land cleanly. (#35676) Thanks @mylukin. - ACPX/runtime: embed the ACP runtime directly in the bundled `acpx` plugin, remove the extra external ACP CLI hop, harden live ACP session binding and reuse, and add a generic `reply_dispatch` hook so bundled plugins like ACPX can own reply interception without hardcoded ACP paths in core auto-reply routing. (#61319) - Agents/progress: add experimental structured plan updates and structured execution item events so compatible UIs can show clearer step-by-step progress during long-running runs. - Providers/Anthropic: remove the Claude CLI backend and setup-token from new onboarding, keep existing configured legacy profiles runnable, and have `openclaw doctor` repair or remove stale `anthropic:claude-cli` state during migration. - Tools/video generation: add bundled xAI (`grok-imagine-video`), Alibaba Model Studio Wan, and Runway video providers, plus live-test/default model wiring for all three. - Memory/search: add Amazon Bedrock embeddings for Titan, Cohere, Nova, and TwelveLabs models, with AWS credential-chain auto-detection for `provider: "auto"` and provider-specific dimension controls. Thanks @wirjo. - Providers/Amazon Bedrock Mantle: generate bearer tokens from the AWS credential chain so Mantle auto-discovery can use IAM auth without manually exporting `AWS_BEARER_TOKEN_BEDROCK`. Thanks @wirjo. - Memory/dreaming (experimental): add weighted short-term recall promotion, a `/dreaming` command, Dreams UI, multilingual conceptual tagging, and doctor/status repair support, while refactoring dreaming from competing modes into three cooperative phases (light, deep, REM) with independent schedules and recovery behavior so durable memory promotion can run in the background with less manual setup. (#60569, #60697) Thanks @vignesh07. - Memory/dreaming: add configurable aging controls (`recencyHalfLifeDays`, `maxAgeDays`) plus optional verbose logging so operators can tune recall decay and inspect promotion decisions more easily. - Memory/dreaming: add REM preview tooling (`openclaw memory rem-harness`, `promote-explain`), surface possible lasting truths during REM staging, and make deep promotion replay-safe so reruns reconcile instead of duplicating `MEMORY.md` entries. - Memory/dreaming: write dreaming trail content to top-level `dreams.md` instead of daily memory notes, update `/dreaming` help text to point there, and keep `dreams.md` available for explicit reads without pulling it into default recall. Thanks @davemorin. - Memory/dreaming: add the Dream Diary surface in Dreams, simplify user-facing dreaming config to `enabled` plus optional `frequency`, treat phases as implementation detail in docs/UI, and keep the lobster animation visible above diary content. Thanks @vignesh07. - Prompt caching: keep prompt prefixes more reusable across transport fallback, deterministic MCP tool ordering, compaction, embedded image history, normalized system-prompt fingerprints, `openclaw status --verbose` cache diagnostics, and the removal of duplicate in-band tool inventories from agent system prompts so follow-up turns hit cache more reliably. (#58036, #58037, #58038, #59054, #60603, #60691) Thanks @bcherny and @vincentkoc. - Agents/cache: diagnostics: add prompt-cache break diagnostics, trace live cache scenarios through embedded runner paths, and show cache reuse explicitly in `openclaw status --verbose`. Thanks @vincentkoc. - Agents/cache: stabilize cache-relevant system prompt fingerprints by normalizing equivalent structured prompt whitespace, line endings, hook-added system context, and runtime capability ordering so semantically unchanged prompts reuse KV/cache more reliably. Thanks @vincentkoc. - Agents/tool prompts: remove the duplicate in-band tool inventory from agent system prompts so tool-calling models rely on the structured tool definitions as the single source of truth, improving prompt stability and reducing stale tool guidance. - Config/schema: enrich the exported `openclaw config schema` JSON Schema with field titles and descriptions so editors, agents, and other schema consumers receive the same config help metadata. (#60067) Thanks @solavrc. - Providers/CLI: remove bundled CLI text-provider backends and the `agents.defaults.cliBackends` surface, while keeping ACP harness sessions and Gemini media understanding on the native bundled providers. - Matrix/exec approvals: clarify unavailable-approval replies so Matrix no longer claims chat approvals are unsupported when native exec approvals are merely unconfigured. (#61424) Thanks @gumadeiras. - Docs/IRC: replace public IRC hostname examples with `irc.example.com` and recommend private servers for bot coordination while listing common public networks for intentional use. - Memory/dreaming: group nearby daily-note lines into short coherent chunks before staging them for dreaming, so one-off context from recent notes reaches REM/deep with better evidence and less line-level noise. - Memory/dreaming: drop generic date/day headings from daily-note chunk prefixes while keeping meaningful section labels, so staged snippets stay cleaner and more reusable. (#61597) Thanks @mbelinky. - Plugins/Lobster: run bundled Lobster workflows in process instead of spawning the external CLI, reducing transport overhead and unblocking native runtime integration. (#61523) Thanks @mbelinky. - Plugins/Lobster: harden managed resume validation so invalid TaskFlow resume calls fail earlier, and memoize embedded runtime loading per runner while keeping failed loads retryable. (#61566) Thanks @mbelinky. ### Fixes - Security: preserve restrictive plugin-only tool allowlists, require owner access for `/allowlist add` and `/allowlist remove`, fail closed when `before_tool_call` hooks crash, block browser SSRF redirect bypasses earlier, and keep non-interactive auth-choice inference scoped to bundled and already-trusted plugins. (#58476, #59836, #59822, #58771, #59120) Thanks @eleqtrizit and @pgondhi987. - Providers/OpenAI: make GPT-5 and Codex runs act sooner with lower-verbosity defaults, visible progress during tool work, and a one-shot retry when a turn only narrates the plan instead of taking action. - Providers/OpenAI and reply delivery: preserve native `reasoning.effort: "none"` and strict schemas where supported, add GPT-5.4 assistant `phase` metadata across replay and the Gateway `/v1/responses` layer, and keep commentary buffered until `final_answer` so web chat, session previews, embedded replies, and Telegram partials stop leaking planning text. Fixes #59150, #59643, #61282. - Telegram: fix current-model checks in the model picker, HTML-format non-default `/model` confirmations, explicit topic replies, persisted reaction ownership across restarts, caption-media placeholder and `file_id` preservation on download failure, and upgraded-install inbound image reads. (#60384, #60042, #59634, #59207, #59948, #59971) Thanks @sfuminya, @GitZhangChi, @dashhuang, @samzong, @v1p0r, and @neeravmakwana. - Telegram: restore DM voice-note preflight transcription so direct-message audio stops arriving as raw `<media:audio>` placeholders. (#61008) Thanks @manueltarouca. - Telegram/reasoning: only create a Telegram reasoning preview lane when the session is explicitly `reasoning:stream`, so hidden `<think>` traces from streamed replies stop surfacing as chat previews on normal sessions. Thanks @vincentkoc. - Telegram/native command menu: trim long menu descriptions before dropping commands so sub-100 command sets can still fit Telegram's payload budget and keep more `/` entries visible. (#61129) Thanks @neeravmakwana. - Discord: keep REST, webhook, and monitor traffic on the configured proxy, preserve component-only media sends, honor `@everyone` and `@here` mention gates, keep ACK reactions on the active account, and split voice connect/playback timeouts so auto-join is more reliable. (#57465, #60361, #60345) Thanks @geekhuashan. - Discord/reply tags: strip leaked `[[reply_to_current]]` control tags from preview text and honor explicit reply-tag threading during final delivery, so Discord replies stay attached to the triggering message instead of printing reply metadata into chat. - Discord/replies: replace the unshipped `replyToOnlyWhenBatched` flag with `replyToMode: "batched"` so native reply references only attach on debounced multi-message turns while explicit reply tags still work. - Discord/image generation: include the real generated `MEDIA:` paths in tool output, avoid duplicate plain-output media requeueing, and persist volatile workspace-generated media into durable outbound media before final reply delivery so generated image replies stop pointing at missing local files. - Slack: route live DM replies back to the concrete inbound DM channel while keeping persisted routing metadata user-scoped, so normal assistant replies stop disappearing when pairing and system messages still arrive. (#59030) Thanks @afurm. - WhatsApp: restore `channels.whatsapp.blockStreaming` and reset watchdog timeouts after reconnect so quiet chats stop falling into reconnect loops. (#60007, #60069) Thanks @MonkeyLeeT and @mcaxtr. - Android/Talk Mode: cancel in-flight `talk.speak` playback when speech is explicitly stopped, and restore spoken replies on both node-scoped and gateway-backed sessions by keeping reply routing and embedded transport overrides aligned with the current playback path. (#60306, #61164, #61214) - Voice-call/OpenAI: pass full plugin config into realtime transcription provider resolution so streaming calls can discover the bundled OpenAI realtime transcription provider again. Fixes #60936. Thanks @sliekens and @vincentkoc. - Matrix/exec approvals: anchor seeded approval reactions to the primary Matrix prompt event, resolve them from event metadata instead of prompt text, and clean up chunked approval prompts correctly. (#60931) Thanks @gumadeiras. - Matrix: recover more reliably when secret storage or recovery keys are missing by recreating secret storage during repair and backup reset, hold crypto snapshot locks during persistence, and surface explicit too-large attachment markers. (#59846, #59851, #60599, #60289) Thanks @al3mart, @emonty, and @efe-arv. - Matrix/DM sessions: add `channels.matrix.dm.sessionScope`, shared-session collision notices, and aligned outbound session reuse so separate Matrix DM rooms can keep distinct context when configured. (#61373) Thanks @gumadeiras. - Matrix: move legacy top-level `avatarUrl` into the default account during multi-account promotion and keep env-backed account setup avatar config persisted. (#61437) Thanks @gumadeiras. - MS Teams: download inline DM images via Graph API and preserve channel reply threading in proactive fallback. (#52212, #55198) Thanks @Ted-developer and @hyojin. - MS Teams: replace the deprecated Teams SDK HttpPlugin stub with `httpServerAdapter` so recurring gateway deprecation warnings stop firing and the Express 5 compatibility workaround stays on the supported SDK path. (#60939) Thanks @coolramukaka-sys. - Control UI/chat: add a per-session thinking-level picker in the chat header and mobile chat settings, and keep the browser bundle on UI-local thinking/session-key helpers so Safari no longer crashes on Node-only imports before rendering chat controls. - Sandbox/SSH: reject hardlinked files during cross-device rename fallback so EXDEV file copies preserve the same pinned file-boundary checks as direct reads. - Control UI: keep Stop visible during tool-only execution, preserve pending-send busy state, and clear stale ClawHub search results as soon as the query changes. (#54528, #59800, #60267) Thanks @chziyue and @frankekn. - Control UI/avatar: honor `ui.assistant.avatar` when serving `/avatar/:agentId` so Appearance UI avatar paths stop falling back to initials placeholders. (#60778) Thanks @hannasdev. - Control UI/cron: highlight the Cron refresh button while refresh is in flight so the page's loading state stays visible even when prior data remains on screen. (#60394) Thanks @coder-zhuzm. - Control UI/Overview: prevent gateway access token/password visibility toggle buttons from overlapping their inputs at narrow widths. (#56924) Thanks @bbddbb1. - Auto-reply: unify reply lifecycle ownership across preflight compaction, session rotation, CLI-backed runs, and gateway restart handling so `/stop` and same-session overlap checks target the right active turn and restart-interrupted turns return the restart notice instead of being silently dropped. (#61267) Thanks @dutifulbob. - Reply delivery: prevent duplicate block replies on `text_end` channels so providers that emit explicit text-end boundaries no longer double-send the same final message. (#61530) - Gateway/startup: default `gateway.mode` to `local` when unset, detect PID recycling in gateway lock files on Windows and macOS, and show startup progress so healthy restarts stop getting blocked by stale locks. (#54801, #60085, #59843) Thanks @BradGroux and @TonyDerek-dot. - Gateway/macOS: let launchd `KeepAlive` own in-process gateway restarts again, adding a short supervised-exit delay so rapid restarts avoid launchd crash-loop unloads while `openclaw gateway restart` still reports real LaunchAgent errors synchronously. - Gateway/macOS: re-bootstrap the LaunchAgent if `launchctl kickstart -k` unloads it during restart so failed restarts do not leave the gateway unmanaged until manual repair. - Gateway/macOS: recover installed-but-unloaded LaunchAgents during `openclaw gateway start` and `restart`, while still preferring live unmanaged gateways during restart recovery. (#43766) Thanks @HenryC-3. - Gateway/Windows scheduled tasks: preserve Task Scheduler settings on reinstall, fail loudly when `/Run` does not start, and report fast failed restarts accurately instead of pretending they timed out after 60 seconds. (#59335) Thanks @tmimmanuel. - Windows/restart: fall back to the installed Startup-entry launcher when the scheduled task was never registered, so `/restart` can relaunch the gateway on Windows setups where `schtasks` install fell back during onboarding. (#58943) Thanks @imechZhangLY. - Windows/restart: clean up stale gateway listeners before Windows self-restart and treat listener and argv probe failures as inconclusive, so scheduled-task relaunch no longer falls into an `EADDRINUSE` retry loop. (#60480) Thanks @arifahmedjoy. - Update/npm: prefer the npm binary that owns the installed global OpenClaw prefix so mixed Homebrew-plus-nvm setups update the right install. (#60153) Thanks @jayeshp19. - Agents/music and video generation: add `tools.media.asyncCompletion.directSend` as an opt-in direct-delivery path for finished async media tasks, while keeping the legacy requester-session wake/model-delivery flow as the default. - CLI/skills JSON: route `skills list --json`, `skills info --json`, and `skills check --json` output to stdout instead of stderr so machine-readable consumers receive JSON on the expected stream again. (#60914; fixes #57599; landed from contributor PR #57611 by @Aftabbs) Thanks @Aftabbs. - CLI/Commander: preserve Commander-computed exit codes for argument and help-error paths, and cover the user-argv parse mode in the regression tests so invalid CLI invocations no longer report success when exits are intercepted. (#60923) Thanks @Linux2010. - Cron: replay interrupted recurring jobs on the first gateway restart instead of waiting for a second restart. (#60583) Thanks @joelnishanth. - Cron: send failure notifications through the job's primary delivery channel using the same session context as successful delivery when no explicit `failureDestination` is configured. (#60622) Thanks @artwalker. - Exec/remote skills: stop advertising `exec host=node` when the current exec policy cannot route to a node, and clarify blocked exec-host override errors with both the requested host and allowed config path. - Agents/Claude CLI/security: clear inherited Claude Code config-root and plugin-root env overrides like `CLAUDE_CONFIG_DIR` and `CLAUDE_CODE_PLUGIN_*`, so OpenClaw-launched Claude CLI runs cannot be silently pointed at an alternate Claude config/plugin tree with different hooks, plugins, or auth context. Thanks @vincentkoc. - Agents/Claude CLI/security: clear inherited Claude Code provider-routing and managed-auth env overrides, and mark OpenClaw-launched Claude CLI runs as host-managed, so Claude CLI backdoor sessions cannot be silently redirected to proxy, Bedrock, Vertex, Foundry, or parent-managed token contexts. Thanks @vincentkoc. - Agents/Claude CLI/security: force host-managed Claude CLI backdoor runs to `--setting-sources user`, even under custom backend arg overrides, so repo-local `.claude` project/local settings, hooks, and plugin discovery do not silently execute inside non-interactive OpenClaw sessions. Thanks @vincentkoc. - Agents/Claude CLI: treat malformed bare `--permission-mode` backend overrides as missing and fail safe back to `bypassPermissions`, so custom `cliBackends.claude-cli.args` security config cannot accidentally consume the next flag as a bogus permission mode. Thanks @vincentkoc. - Gateway/device pairing: require non-admin paired-device sessions to manage only their own device for token rotate/revoke and paired-device removal, blocking cross-device token theft inside pairing-scoped sessions. (#50627) Thanks @coygeek. - Gateway/plugin routes: keep gateway-auth plugin runtime routes on write-only fallback scopes unless a trusted-proxy caller explicitly declares narrower `x-openclaw-scopes`, so plugin HTTP handlers no longer mint admin-level runtime scopes on missing or untrusted HTTP scope headers. (#59815) Thanks @pgondhi987. - Build/types: fix the Node `createRequire(...)` helper typing so provider-runtime lazy loads compile cleanly again and `pnpm build` no longer fails in the Pi embedded provider error-pattern path. - Gateway/security: scope loopback browser-origin auth throttling by normalized origin so one localhost Control UI tab cannot lock out a different localhost browser origin after repeated auth failures. - Gateway/auth: serialize async shared-secret auth attempts per client so concurrent Tailscale-capable failures cannot overrun the intended auth rate-limit budget. Thanks @Telecaster2147. - Device pairing/security: keep non-operator device scope checks bound to the requested role prefix so bootstrap verification cannot redeem `operator.*` scopes through `node` auth. (#57258) Thanks @jlapenna. - Device pairing: reject rotating device tokens into roles that were never approved during pairing, and keep reconnect role checks bounded to the paired device's approved role set. (#60462) Thanks @eleqtrizit. - Gateway/device auth: reuse cached device-token scopes only for cached-token reconnects, while keeping explicit `deviceToken` scope requests and empty-cache fallbacks intact so reconnects preserve `operator.read` without breaking explicit auth flows. (#46032) Thanks @caicongyang. - Mobile pairing/security: fail closed for internal `/pair` setup-code issuance, cleanup, and approval paths when gateway pairing scopes are missing, and keep approval-time requested-scope enforcement on the internal command path. (#55996) Thanks @coygeek. - Mobile pairing/bootstrap: keep QR bootstrap handoff tokens bounded to the mobile-safe contract so node handoff stays unscoped and operator handoff drops mixed `node.*`, `operator.admin`, and `operator.pairing` scopes. - Mobile pairing/Android: tighten secure endpoint handling so Tailscale and public remote setup reject cleartext endpoints, private LAN pairing still works, merged-role approvals mint both node and operator device tokens, and bootstrap tokens survive node auto-pair until operator approval finishes. (#60128, #60208, #60221) Thanks @obviyus. - Android/canvas security: require exact normalized A2UI URL matches before forwarding canvas bridge actions, rejecting query mismatches and descendant paths while still allowing fragment-only A2UI navigation. - Synology Chat/security: default low-level HTTPS helper TLS verification to on so helper/API defaults match the shipped safe account default, and only explicit `allowInsecureSsl: true` opts out. - Synology Chat/security: route webhook token comparison through the shared constant-time secret helper for consistency with other bundled plugins. - Plugins/marketplace: block remote marketplace symlink escapes without breaking ordinary local marketplace install paths. (#60556) Thanks @eleqtrizit. - Telegram/local Bot API: honor `channels.telegram.apiRoot` for buffered media downloads, add `channels.telegram.network.dangerouslyAllowPrivateNetwork` for trusted fake-IP setups, and require `channels.telegram.trustedLocalFileRoots` before reading absolute Bot API `file_path` values. (#59544, #60705) Thanks @SARAMALI15792 and @obviyus. - Outbound/sanitizer: strip leaked `<tool_call>`, `<function_calls>`, and model special tokens from shared user-visible assistant text, including truncated tool-call streams, so internal scaffolding no longer bleeds into replies across surfaces. (#60619) Thanks @oliviareid-svg. - Agents/errors: surface an explicit disk-full message when local session or transcript writes fail with `ENOSPC`/`disk full`, so those runs stop degrading into opaque `NO_REPLY`-style failures. Thanks @vincentkoc. - Exec approvals: remove heuristic command-obfuscation gating from host exec so gateway and node runs rely on explicit policy, allowlist, and strict inline-eval rules only. - Agents/tool results: cap live tool-result persistence and overflow-recovery truncation at 40k characters so oversized tool output stays bounded without discarding recent context entirely. - Discord/video replies: split text-plus-video deliveries into a text reply followed by a media-only send, and let live provider auth checks honor manifest-declared API key env vars like `MODELSTUDIO_API_KEY`. - Config/All Settings: keep the raw config view intact when sensitive fields are blank instead of corrupting or dropping the rendered snapshot. (#28214) Thanks @solodmd. - Plugin SDK/facades: back-fill bundled plugin facade sentinels before plugin-id tracking re-enters config loading, so CLI/provider startup no longer crashes with `shouldNormalizeGoogleProviderConfig is not a function` or other empty-facade reads during bundled plugin re-entry. Thanks @adam91holt. - Plugins/facades: back-fill facade sentinels before tracked-plugin resolution re-enters config loading, so facade exports stay defined during circular provider normalization. (#61180) Thanks @adam91holt. - QA lab: restore typed mock OpenAI gateway config wiring so QA-lab config helpers compile cleanly again and `pnpm check` / `pnpm build` stay green. - Discord/image generation: include the real generated `MEDIA:` paths in tool output and avoid duplicate plain-output media requeueing so Discord image replies stop pointing at missing local files. - Slack: route live DM replies back to the concrete inbound DM channel while keeping persisted routing metadata user-scoped, so normal assistant replies stop disappearing when pairing and system messages still arrive. (#59030) Thanks @afurm. - Discord/reply tags: strip leaked `[[reply_to_current]]` control tags from preview text and honor explicit reply-tag threading during final delivery, so Discord replies stay attached to the triggering message instead of printing reply metadata into chat. - Telegram: fix current-model checks in the model picker, HTML-format non-default `/model` confirmations, explicit topic replies, persisted reaction ownership across restarts, caption-media placeholder and `file_id` preservation on download failure, and upgraded-install inbound image reads. (#60384, #60042, #59634, #59207, #59948, #59971) Thanks @sfuminya, @GitZhangChi, @dashhuang, @samzong, @v1p0r, and @neeravmakwana. - Telegram: restore DM voice-note preflight transcription so direct-message audio stops arriving as raw `<media:audio>` placeholders. (#61008) Thanks @manueltarouca. - Telegram/reasoning: only create a Telegram reasoning preview lane when the session is explicitly `reasoning:stream`, so hidden `<think>` traces from streamed replies stop surfacing as chat previews on normal sessions. Thanks @vincentkoc. - Telegram/native command menu: trim long menu descriptions before dropping commands so sub-100 command sets can still fit Telegram's payload budget and keep more `/` entries visible. (#61129) Thanks @neeravmakwana. - Feishu/reasoning: only expose streamed reasoning previews when the session is explicitly `reasoning:stream`, so hidden reasoning traces do not surface on normal streaming sessions. Thanks @vincentkoc. - Discord: keep REST, webhook, and monitor traffic on the configured proxy, preserve component-only media sends, honor `@everyone` and `@here` mention gates, keep ACK reactions on the active account, and split voice connect/playback timeouts so auto-join is more reliable. (#57465, #60361, #60345) Thanks @geekhuashan. - WhatsApp: restore `channels.whatsapp.blockStreaming` and reset watchdog timeouts after reconnect so quiet chats stop falling into reconnect loops. (#60007, #60069) Thanks @MonkeyLeeT and @mcaxtr. - Memory: keep `memory-core` builtin embedding registration on the already-registered path so selecting `memory-core` no longer recurses through plugin discovery and crashes during startup. (#61402) Thanks @ngutman. - Agents/tool results: keep large `read` outputs visible longer, preserve the latest `read` output when older tool output can absorb the overflow budget, and fall back to Pi's normal overflow compaction/retry path before replacing a fresh `read` with a compacted stub. Thanks @vincentkoc. - Memory/QMD: prefer modern `qmd collection add --glob`, accept newer single-line JSON hit metadata while keeping legacy line fields, refresh QMD docs/doctor install guidance and model-override guidance, and keep older QMD releases working. Thanks @vincentkoc. - MS Teams: download inline DM images via Graph API and preserve channel reply threading in proactive fallback. (#52212, #55198) Thanks @Ted-developer and @hyojin. - MS Teams: replace the deprecated Teams SDK HttpPlugin stub with `httpServerAdapter` so recurring gateway deprecation warnings stop firing and the Express 5 compatibility workaround stays on the supported SDK path. (#60939) Thanks @coolramukaka-sys. - Matrix/exec approvals: anchor seeded approval reactions to the primary Matrix prompt event, resolve them from event metadata instead of prompt text, and clean up chunked approval prompts correctly. (#60931) Thanks @gumadeiras. - Matrix: recover more reliably when secret storage or recovery keys are missing by recreating secret storage during repair and backup reset, hold crypto snapshot locks during persistence, and surface explicit too-large attachment markers. (#59846, #59851, #60599, #60289) Thanks @al3mart, @emonty, and @efe-arv. - Android/Talk Mode: cancel in-flight `talk.speak` playback when speech is explicitly stopped, so stale replies stop starting after barge-in or manual stop. (#61164) Thanks @obviyus. - Android/Talk Mode: restore spoken assistant replies on node-scoped sessions by keeping reply routing synced to the resolved node session key and pausing mic capture during reply playback. (#60306) Thanks @MKV21. - Android/Talk Mode: restore voice replies on gateway-backed talk mode sessions by updating embedded runner transport overrides to the current agent transport API. (#61214) Thanks @obviyus. - Voice-call/OpenAI: pass full plugin config into realtime transcription provider resolution so streaming calls can discover the bundled OpenAI realtime transcription provider again. Fixes #60936. Thanks @sliekens and @vincentkoc. - Control UI/chat: add a per-session thinking-level picker in the chat header and mobile chat settings, and keep the browser bundle on UI-local thinking/session-key helpers so Safari no longer crashes on Node-only imports before rendering chat controls. - Control UI: keep Stop visible during tool-only execution, preserve pending-send busy state, and clear stale ClawHub search results as soon as the query changes. (#54528, #59800, #60267) Thanks @chziyue and @frankekn. - Control UI/avatar: honor `ui.assistant.avatar` when serving `/avatar/:agentId` so Appearance UI avatar paths stop falling back to initials placeholders. (#60778) Thanks @hannasdev. - Control UI/cron: highlight the Cron refresh button while refresh is in flight so the page's loading state stays visible even when prior data remains on screen. (#60394) Thanks @coder-zhuzm. - Control UI/Overview: prevent gateway access token/password visibility toggle buttons from overlapping their inputs at narrow widths. (#56924) Thanks @bbddbb1. - CLI/skills JSON: route `skills list --json`, `skills info --json`, and `skills check --json` output to stdout instead of stderr so machine-readable consumers receive JSON on the expected stream again. (#60914; fixes #57599; landed from contributor PR #57611 by @Aftabbs) Thanks @Aftabbs. - CLI/Commander: preserve Commander-computed exit codes for argument and help-error paths, and cover the user-argv parse mode in the regression tests so invalid CLI invocations no longer report success when exits are intercepted. (#60923) Thanks @Linux2010. - Cron: replay interrupted recurring jobs on the first gateway restart instead of waiting for a second restart. (#60583) Thanks @joelnishanth. - Cron: send failure notifications through the job's primary delivery channel using the same session context as successful delivery when no explicit `failureDestination` is configured. (#60622) Thanks @artwalker. - Live model switching: only treat explicit user-driven model changes as pending live switches, so fallback rotation, heartbeat overrides, and compaction no longer trip `LiveSessionModelSwitchError` before making an API call. (#60266) Thanks @kiranvk-2011. - Exec approvals: reuse durable exact-command `allow-always` approvals in allowlist mode so identical reruns stop prompting, and tighten Windows interpreter/path approval handling so wrapper and malformed-path cases fail closed more consistently. (#59880, #59780, #58040, #59182) Thanks @luoyanglang, @SnowSky1, and @pgondhi987. - Node exec approvals: keep node-host `system.run` approvals bound to the prepared execution plan across async forwarding, so mutable script operands still get approval-time binding and drift revalidation instead of dropping back to unbound execution. - Agents/exec approvals: let `exec-approvals.json` agent security override stricter gateway tool defaults so approved subagents can use `security: “full”` without falling back to allowlist enforcement again. (#60310) Thanks @lml2468. - Agents/exec: restore `host=node` routing for node-pinned and `host=auto` sessions, while still blocking sandboxed `auto` sessions from jumping to gateway. (#60788) Thanks @openperf. - Exec/heartbeat: use the canonical `exec-event` wake reason for `notifyOnExit` so background exec completions still trigger follow-up turns when `HEARTBEAT.md` is empty or comments-only. (#41479) Thanks @rstar327. - Heartbeat: skip wake delivery when the target session lane is already busy so the pending event is retried instead of getting drained too early. (#40526) Thanks @lucky7323. - Group chats/agent prompts: tell models to minimize empty lines and use normal chat-style spacing so group replies avoid document-style blank-line formatting. - Providers/OpenAI GPT: treat short approval turns like `ok do it` and `go ahead` as immediate action turns, and trim overly memo-like GPT-5 chat confirmations so OpenAI replies stay shorter and more conversational by default. - Providers/OpenAI Codex: split native `contextWindow` from runtime `contextTokens`, keep the default effective cap at `272000`, and expose a per-model `contextTokens` override on `models.providers.*.models[]`. - Providers/OpenAI-compatible WS: compute fallback token totals from normalized usage when providers omit or zero `total_tokens`, so DashScope-compatible sessions stop storing zero totals after alias normalization. (#54940) Thanks @lyfuci. - Agents/OpenAI: mark Claude-compatible file tool schemas as `additionalProperties: false` so direct OpenAI GPT-5 routes stop rejecting the `read` tool with invalid strict-schema errors. - Agents/OpenAI: fall back to `strict: false` for native OpenAI tool calls when a tool schema is not strict-compatible, and normalize empty-object tool schemas to include `required: []`, so direct GPT-5 routes stop failing with invalid strict-schema errors like missing `path` in `required`. - Agents/GPT: add explicit work-item lifecycle events for embedded runs, use them to surface real progress more reliably, and stop counting tool-started turns as planning-only retries. - Plugins/OpenAI: enable `gpt-image-1` reference-image edits through `/images/edits` multipart uploads, and stop inferring unsupported resolution overrides when no explicit `size` or `resolution` is provided. - Agents/replay: remove the malformed assistant-content canonicalization repair from replay history sanitization instead of extending that legacy repair path into replay validation. - Plugins/OpenAI: tune the OpenAI prompt overlay for live-chat cadence so GPT replies stay shorter, more human, and less wall-of-text by default. - Providers/compat: stop forcing OpenAI-only defaults on proxy and custom OpenAI-compatible routes, preserve native vendor-specific reasoning/tool/streaming behavior across Anthropic-compatible, Moonshot, Mistral, ModelStudio, OpenRouter, xAI, and Z.ai endpoints, and route GitHub Copilot Claude models through Anthropic Messages instead of OpenAI Responses. - Providers/GitHub Copilot: send IDE identity headers on runtime model requests and GitHub token exchange so IDE-authenticated Copilot runs stop failing with missing `Editor-Version`. (#60641) Thanks @VACInc and @vincentkoc. - Providers/OpenRouter failover: classify `403 “Key limit exceeded”` spending-limit responses as billing so model fallback continues instead of stopping on generic auth. (#59892) Thanks @rockcent. - Providers/Anthropic: keep `claude-cli/*` auth on live Claude CLI credentials at runtime, avoid persisting stale bearer-token profiles, and suppress macOS Keychain prompts during non-interactive Claude CLI setup. (#61234) Thanks @darkamenosa. - Providers/Anthropic: when Claude CLI auth becomes the default, write a real `claude-cli` auth profile so local and gateway agent runs can use Claude CLI immediately without missing-API-key failures. Thanks @vincentkoc. - Providers/Anthropic Vertex: honor `cacheRetention: “long”` with the real 1-hour prompt-cache TTL on Vertex AI endpoints, and default `anthropic-vertex` cache retention like direct Anthropic. (#60888) Thanks @affsantos. - Agents/Anthropic: preserve native `toolu_*` replay ids on direct Anthropic and Anthropic Vertex paths so cache-sensitive history stops rewriting known-valid Anthropic tool-use ids. (#52612) - Providers/Google: add model-level `cacheRetention` support for direct Gemini system prompts by creating, reusing, and refreshing `cachedContents` automatically on Google AI Studio runs. (#51372) Thanks @rafaelmariano-glitch. - Google Gemini CLI auth: detect bundled npm installs by scanning packaged bundle files for the Gemini OAuth client config, so `npm install -g @google/gemini-cli` layouts work again. (#60486) Thanks @wzfmini01. - Google Gemini CLI auth: detect personal OAuth mode from local Gemini settings and skip Code Assist project discovery for those logins, so personal Google accounts stop failing with `loadCodeAssist 400 Bad Request`. (#49226) Thanks @bobworrall. - Google Gemini CLI auth: improve OAuth credential discovery across Windows nvm and Homebrew libexec installs, and align Code Assist metadata so Gemini login stops failing on packaged CLI layouts. (#40729) Thanks @hughcube. - Google Gemini CLI models: add forward-compat support for stable `gemini-2.5-*` model ids by letting the bundled CLI provider clone them from Google templates, so `gemini-2.5-flash-lite` and related configured models stop showing up as missing. (#35274) Thanks @mySebbe. - Google image generation: disable pinned DNS for Gemini image requests and honor explicit `pinDns` overrides in shared provider HTTP helpers so proxy-backed image generation works again. (#59873) Thanks @luoyanglang. - Providers/Microsoft Foundry: preserve explicit image capability on normalized Foundry deployments, repair stale GPT/o-series text-only model metadata across gateway and runtime paths, and keep unknown fallback models from borrowing unrelated image support. - Providers/Model Studio: preserve native streaming usage reporting for DashScope-compatible endpoints even when they are configured under a generic provider key, so streamed token totals stop sticking at zero. (#52395) Thanks @IVY-AI-gif. - Providers/Z.AI: preserve explicitly registered `glm-5-*` variants like `glm-5-turbo` instead of intercepting them with the generic GLM-5 forward-compat shim. (#48185) Thanks @haoyu-haoyu. - Amazon Bedrock/aws-sdk auth: stop injecting the fake `AWS_PROFILE` apiKey marker when no AWS auth env vars exist, so instance-role and other default-chain setups keep working without poisoning provider config. (#61194) Thanks @wirjo. - Agents/Kimi tool-call repair: preserve tool arguments that were already present on streamed tool calls when later malformed deltas fail reevaluation, while still dropping stale repair-only state before `toolcall_end`. - Plugins/Kimi Coding: parse tagged tool calls and keep Anthropic-native tool payloads so Kimi coding endpoints execute tools instead of echoing raw markup. (#60051, #60391) Thanks @obviyus and @Eric-Guo. - Media understanding: auto-register image-capable config providers for vision routing, so custom GLM-style provider ids with image models stop failing with “no media-understanding provider registered”. (#51418) Thanks @xydt-610. - Plugins/media understanding: enable bundled Groq and Deepgram providers by default so configured transcription models work without extra plugin activation config. (#59982) Thanks @yxjsxy. - MiniMax/pricing: keep bundled MiniMax highspeed pricing distinct in provider catalogs and preserve the lower M2.5 cache-read pricing when onboarding older MiniMax models. (#54214) Thanks @octo-patch. - MiniMax: advertise image input on bundled `MiniMax-M2.7` and `MiniMax-M2.7-highspeed` model definitions so image-capable flows can route through the M2.7 family correctly. (#54843) Thanks @MerlinMiao88888888. - Models/MiniMax: honor `MINIMAX_API_HOST` for implicit bundled MiniMax provider catalogs so China-hosted API-key setups pick `api.minimaxi.com/anthropic` without manual provider config. (#34524) Thanks @caiqinghua. - Usage/MiniMax: invert remaining-style `usage_percent` fields when MiniMax reports only remaining percentage data, so usage bars stop showing nearly-full remaining quota as nearly-exhausted usage. (#60254) Thanks @jwchmodx. - Usage/MiniMax: let usage snapshots treat `minimax-portal` and MiniMax CN aliases as the same MiniMax quota surface, and prefer stored MiniMax OAuth before falling back to Coding Plan keys. - Usage/MiniMax: prefer the chat-model `model_remains` entry and derive Coding Plan window labels from MiniMax interval timestamps so MiniMax usage snapshots stop picking zero-budget media rows and misreporting 4h windows as `5h`. (#52349) Thanks @IVY-AI-gif. - Model picker/providers: treat bundled BytePlus and Volcengine plan aliases as their native providers during setup, and expose their bundled standard/coding catalogs before auth so setup can suggest the right models. (#58819) Thanks @Luckymingxuan. - Tools/web_search (Kimi): when `tools.web.search.kimi.baseUrl` is unset, inherit native Moonshot chat `baseUrl` (`.ai` / `.cn`) so China console keys authenticate on the same host as chat. Fixes #44851. (#56769) Thanks @tonga54. - Agents/Claude CLI: keep non-interactive `--permission-mode bypassPermissions` when custom `cliBackends.claude-cli.args` override defaults, including fallback resolution before the runtime plugin registry is active, so cron and heartbeat Claude CLI runs do not regress to interactive approval mode. (#61114) Thanks @cathrynlavery and @thewilloftheshadow. - Agents/Claude CLI: persist explicit `openclaw agent --session-id` runs under a stable session key so follow-ups can reuse the stored CLI binding and resume the same underlying Claude session. - Agents/Claude CLI: persist routed Claude session bindings, rotate them on `/new` and `/reset`, and keep live Claude CLI model switches moving across the configured Claude family so resumed sessions follow the real active thread and model. Thanks @vincentkoc. - Agents/CLI backends: invalidate stored CLI session reuse when local CLI login state or the selected auth profile credential changes, so relogin and token rotation stop resuming stale sessions. - Agents/Claude CLI/images: reuse stable hydrated image file paths and preserve shared media extensions like HEIC when passing image refs to local CLI runs, so Claude CLI image prompts stop thrashing KV cache prefixes and oddball image formats do not fall back to `.bin`. Thanks @vincentkoc. - Agents/compaction: keep assistant tool calls and displaced tool results in the same compaction chunk so strict summarization providers stop rejecting orphaned tool pairs. (#58849) Thanks @openperf. - Agents/failover: scope Anthropic `An unknown error occurred` failover matching by provider so generic internal unknown-error text no longer triggers retryable timeout fallback. (#59325) Thanks @aaron-he-zhu. - Agents/subagents: honor allowlist validation, auth-profile handoff, and session override state when a subagent retries after `LiveSessionModelSwitchError`. (#58178) Thanks @openperf. - Agents/runtime: make default subagent allowlists, inherited skills/workspaces, and duplicate session-id resolution behave more predictably, and include value-shape hints in missing-parameter tool errors. (#59944, #59992, #59858, #55317) Thanks @hclsys, @gumadeiras, @joelnishanth, and @priyansh19. - Agents/pairing: merge completion announce delivery context with the requester session fallback so missing `to` still reaches the original channel, and include `operator.talk.secrets` in CLI default operator scopes for node-role device pairing approvals. (#56481) Thanks @maxpetrusenko. - Agents/scheduling: steer background-now work toward automatic completion wake and treat `process` polling as on-demand inspection or intervention instead of default completion handling. (#60877) Thanks @vincentkoc. - Agents/skills: skip `.git` and `node_modules` when mirroring skills into sandbox workspaces so read-only sandboxes do not copy repo history or dependency trees. (#61090) Thanks @joelnishanth. - ACP/agents: inherit the target agent workspace for cross-agent ACP spawns and fall back safely when the inherited workspace no longer exists. (#58438) Thanks @zssggle-rgb. - ACPX/Windows: preserve backslashes and absolute `.exe` paths in Claude CLI parsing, and fail fast on wrapper-script targets with guidance to use `cmd.exe /c`, `powershell.exe -File`, or `node <script>`. (#60689) Thanks @steipete. - Auth/failover: persist selected fallback overrides before retrying, shorten `auth_permanent` lockouts, and refresh websocket/shared-auth sessions only when real auth changes occur so retries and secret rotations behave predictably. (#60404, #60323, #60387) Thanks @extrasmall0 and @mappel-nv. - Gateway/channels: pin the initial startup channel registry before later plugin-registry churn so configured channels stay visible and `channels.status` stops falling back to empty `channelOrder` / `channels` payloads after runtime plugin loads. - Prompt caching: order stable workspace project-context files before `HEARTBEAT.md` and keep `HEARTBEAT.md` below the system-prompt cache boundary so heartbeat churn does not invalidate the stable project-context prefix. (#58979) Thanks @yozu and @vincentkoc. - Prompt caching: route Codex Responses and Anthropic Vertex through boundary-aware cache shaping, and report the actual outbound system prompt in cache traces so cache reuse and misses line up with what providers really receive. Thanks @vincentkoc. - Agents/cache: preserve the full 3-turn prompt-cache image window across tool loops, keep colliding bundled MCP tool definitions deterministic, and reapply Anthropic Vertex cache shaping after payload hook replacements so KV/cache reuse stays stable. Thanks @vincentkoc. - Status/cache: restore `cacheRead` and `cacheWrite` in transcript fallback so `/status` keeps showing cache hit percentages when session logs are the only complete usage source. (#59247) Thanks @stuartsy. - Status/usage: let `/status` and `session_status` fall back to transcript token totals when the session meta store stayed at zero, so LM Studio, Ollama, DashScope, and similar OpenAI-compatible providers stop showing `Context: 0/...`. (#55041) Thanks @jjjojoj. - Mattermost/config schema: accept `groups.*.requireMention` again so existing Mattermost configs no longer fail strict validation after upgrade. (#58271) Thanks @MoerAI. - Doctor/config: compare normalized `talk` configs by deep structural equality instead of key-order-sensitive serialization so `openclaw doctor --fix` stops repeatedly reporting/applying no-op `talk.provider/providers` normalization. (#59911) Thanks @ejames-dev. - Anthropic CLI onboarding: rewrite migrated fallback model refs during non-interactive Claude CLI setup too, so onboarding and scripted setup no longer keep stale `anthropic/*` fallbacks after switching the primary model to `claude-cli/*`. Thanks @vincentkoc. - Models/Anthropic CLI auth: replace migrated `agents.defaults.models` allowlists when `openclaw models auth login --provider anthropic --method cli --set-default` switches to `claude-cli/*`, so stale `anthropic/*` entries do not linger beside the migrated Claude CLI defaults. Thanks @vincentkoc. - Doctor/Claude CLI: add dedicated Claude CLI health checks so `openclaw doctor` can spot missing local installs or broken auth before agent runs fail. Thanks @vincentkoc. - Plugins/auth-choice: apply provider-owned auth config patches without recursively preserving replaced default-model maps, so Anthropic Claude CLI and similar migrations can intentionally swap model allowlists during onboarding and setup instead of accumulating stale entries. Thanks @vincentkoc. - Plugins/onboarding: write dotted plugin uiHint paths like Brave `webSearch.mode` as nested plugin config so `llm-context` setup stops failing validation. (#61159) Thanks @obviyus. - Plugins/install: preserve unsafe override flags across linked plugin and hook-pack probes so local `--link` installs honor the documented override behavior. (#60624) Thanks @JerrettDavis. - Plugins/cache: inherit the active gateway workspace for provider, web-search, and web-fetch snapshot loads when callers omit `workspaceDir`, so compatible plugin registries and snapshot caches stop missing on gateway-owned runtime paths. (#61138) Thanks @jzakirov. - Plugin SDK/context engines: export the missing context-engine result and subagent lifecycle types from `openclaw/plugin-sdk` so context engine plugins can type `ContextEngine` implementations without local workarounds. (#61251) Thanks @DaevMithran. - Tasks/maintenance: reconcile stale cron and chat-backed CLI task rows against live cron-job and agent-run ownership instead of treating any persisted session key as proof that the task is still running. (#60310) Thanks @lml2468. - Plugins: suppress trust-warning noise during non-activating snapshot and CLI metadata loads. (#61427) Thanks @gumadeiras. - Agents/video generation: accept `agents.defaults.videoGenerationModel` in strict config validation and `openclaw config set/get`, so gateways using `video_generate` no longer fail to boot after enabling a video model. - Matrix/streaming: add a quiet preview mode for streamed Matrix replies, keep legacy `partial` preview-first behavior, and finalize quiet media captions correctly so previews stop notifying early without dropping final text semantics. (#61450) Thanks @gumadeiras. - Gateway/shutdown: bound websocket-server shutdown even when no tracked clients remain, so gateway restarts stop hanging until the watchdog kills the process. (#61565) Thanks @mbelinky. - Control UI/multilingual: localize the remaining shared channel, instances, nodes, and gateway-confirmation strings so the dashboard stops mixing translated UI with hardcoded English labels. Thanks @vincentkoc. - Discord/media: raise the default inbound and outbound media cap to `100MB` so Discord matches Telegram more closely and larger attachments stop failing on the old low default. - Matrix: keep direct transport requests on the pinned dispatcher by routing them through undici runtime fetch, so Matrix clients resume syncing on newer runtimes without dropping the validated address binding. (#61595) Thanks @gumadeiras. - Plugins/facades: resolve globally installed bundled-plugin runtime facades from registry roots so bundled channels like LINE still boot when the winning plugin install lives under the global extensions directory with an encoded scoped folder name. (#61297) Thanks @openperf.