OpenClaw Newsletter - 2026-07-31
OpenClaw Newsletter
Friday, July 31, 2026Top Stories
- fix(models): normalize partial cost in generated plugin catalogs to prevent schema validation warnings — Fixes #116116 What Problem This Solves Fixes an issue where operators who configure custom Anthropic model cost overlays with partial pricing (e.g. only input and output, without...
- fix(ui): stop realtime Talk from retaining runaway audio — Related: #116201 What Problem This Solves Browser Realtime Talk could retain an unbounded number of PCM playback sources, decoded samples, and queued playback time when provider output arrived faster...
- fix(media): preserve code whitespace in media replies — Closes #116458 What Problem This Solves Media-bearing assistant replies could lose meaningful code formatting when MEDIA attachment directives or extracted Telegram Markdown images were removed from...
- [Bug]: Context refresh after compaction replays old inbound message_id — missing gateway-level dedup for Telegram — Bug type Behavior bug — stale turn replay after compaction-triggered context refresh in Telegram DM session. Summary When OpenClaw auto-compacts an active Telegram DM session (triggered by context...
- fix(slack): preserve context in group DM thread follow-ups — Fixes #113449. What Problem This Solves Fixes an issue where Slack group-DM users would lose the opening exchange when continuing inside the bot's reply thread. The first threaded follow-up could...
- fix(status): report health, live plugin hooks, and real account availability — Summary - Preserve deep JSON health-probe failures and security audit results instead of returning an apparently healthy status. - Read compatibility notices from genuinely active plugin hook...
- feat(plugins): externalize DuckDuckGo search — What Problem This Solves DuckDuckGo web search is optional, but its plugin currently ships inside every core OpenClaw package. Operators who never use this keyless experimental provider still receive...
- fix(ui): coalesce profile identity refresh — What Problem This Solves Fixes an issue where repeated Profile Refresh activation could fan out concurrent users.self requests, and a same-client reconnect could leave stale identity request...
- fix: pending outbound sends are safely retired on upgrade — Related: #116632 What Problem This Solves Fixes an issue where upgrading across the D4 outbound-queue boundary could retain pending legacy messages through a complex replay migration. That migration...
- fix(mistral): centralize final realtime transcription ownership — Summary - Give Mistral realtime final transcription events a single canonical owner instead of replaying aggregate done.text over already committed segments. - Preserve every legitimate independently...
- [Bug] Reply is silently truncated at [[ when the text contains no matching ]] — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary When an assistant reply contains an unterminated [[ sequence, the delivered reply is truncated at that...
- feat: add per-agent daily model spend alerts — Closes #113481 AI-assisted: this change was implemented with Codex assistance. I reviewed the code and understand the implementation. No agent transcript is included. What Problem This Solves...
- fix(release): skip unavailable Buzz QA runner — What changed - detect whether the selected immutable release ref contains the Buzz QA runner - run Buzz credential validation, private QA build, and live proof only when that runner exists - upload...
- fix(openai): close preconnect realtime sessions — Related: #116201 What Problem This Solves Fixes an issue where closing an OpenAI realtime voice session before its first connection left queued audio and lifecycle state alive. A later explicit...
- fix(talk): diagnose invalid OpenAI realtime auth — SummaryWhat problem does this PR solve?- Talk configured for OpenAI Realtime could surface a raw provider error such as Incorrect API key provided when the available credential was an Azure AI...
- fix(chat): explain chat-only model limits — Closes #91434 What Problem This Solves Fixes an issue where users choosing an LM Studio or other explicitly chat-only model could see the assistant claim file, command, web, or media work happened...
- fix(commands): validate native settings and preserve Telegram ingress outcomes — Related: #116688 Follow-up to #116726 What Problem This Solves Fixes an issue where Telegram, Discord, and Slack users invoking native session-settings commands with malformed or trailing arguments...
- fix(doctor): exclude disabled automation overrides and report SQLite storage — Refs #116568 What Problem This Solves Fixes an issue where operators running openclaw doctor saw inactive automation jobs counted as active model overrides and were directed to a nonexistent retired...
- refactor(reply): simplify task-owned typing refresh — Related: #116695 Follow-up to #116721 What Problem This Solves Keeps the recently landed Telegram task-owned typing fix easy to maintain by removing an unused result and duplicate timer refresh from...
- fix(packaging): keep tracked skills visible without a brittle global ignore list — Summary - Remove the brittle repository-wide .agents/skills/* denylist and its 45 duplicated, manually maintained allowlist exceptions. - Keep generated skill scratch artifacts ignored while leaving...
- docs(hooks): mark pairing hook observation-only — What Problem This Solves Fixes an issue where plugin authors could read the hook catalog as saying channel_pairing_requested accepts an authorization decision, even though the runtime dispatches it...
- feat: add task flow lifecycle observability with tags, hooks, and diagnostic events — Summary Adds comprehensive observability for the task flow lifecycle via diagnostic events, internal hooks, and OTel/Prometheus metrics. Also introduces a tags field on TaskFlowRecord for arbitrary...
- fix(ui): preserve session pages and coordinate scoped gateway refreshes — Summary - Preserve loaded archived/all session pages across matching Gateway refreshes; reset pagination only when the selected agent, filter, client, or generation actually changes. - Route...
- [Bug]: CLI and TUI rerun a full state-database integrity scan after startup migrations completed — Bug type Crash (process/app exits or hangs) Beta release blocker No Summary On OpenClaw 2026.7.2-beta.3, ordinary CLI commands such as openclaw status and openclaw tui can appear hung while...
- fix(browser): retire durable tab rows whose browser never returns — What Problem This Solves Resolves a problem where a user whose browser disappears — closed for good, moved to a different cdpUrl, a laptop that never wakes on the same port — permanently loses the...
- fix: stabilize embedded prompt caching across policy and Responses boundaries — Related: #102175 What Problem This Solves Long-lived embedded sessions could lose prompt-cache reuse when the same session crossed room-event delivery, sender-policy, queue, compaction, recovery, or...
- fix(pairing): ignore wildcard bindings during legacy migration — What Problem This Solves OpenClaw 2026.7.2-beta.5 can enter a gateway crash-restart loop while migrating legacy channel pairing state. A route binding with accountId: "" is a wildcard selector, but...
- test(ui): emulate reduced motion in shared e2e gateway setup — What Adds await page.emulateMedia({ reducedMotion: "reduce" }) to installMockGateway() in ui/src/test-helpers/control-ui-e2e.ts — the shared mocked-Gateway page setup used by 93 of the Control UI E2E...
- feat(apple): inline audio/video players in iOS and macOS chat — What Problem This Solves Apple chat clients can render managed images inline, but managed audio and video artifacts fall back to non-native attachment handling instead of giving users an in-chat...
- feat(android): inline audio/video players with Media3 — What Problem This Solves Android chat attachments currently render audio and video as generic files, so users must leave the conversation to play media. Part of #115590 Why This Change Was Made Add...
- OpenClaw repo stats for 2026-07-31 — 384657 stars, 80846 forks, 3028 contributors
- OpenClaw sponsors (98 total) — 98 sponsors supporting OpenClaw
- fix: accept direct workflow oidc claims (#3348)
- fix: keep inspector target cache in workspace (#3347)
- fix: plugin publishing no longer fails on invalid temp paths (#3344) — fix: keep plugin inspector workspaces writable test: cover inspector temp fallback by platform Co-authored-by: Vincent Koc
- fix: retry transient legacy file downloads (#3346)
- fix: reconstruct large legacy scan packages (#3345)
- fix: normalize nested plugin manifests (#3343)
- fix: normalize BOM-prefixed plugin manifests (#3342)
- fix: normalize BOM-prefixed plugin manifests (#3341)
- fix: pin BOM-compatible plugin inspector (#3339)
- fix: accept BOM-prefixed plugin package metadata (#3338)
- fix: scan historical legacy package paths (#3337)
- fix: sanitize nightly scan fixture ids (#3336)
- fix: ignore legacy zip pax metadata (#3335)
- fix: label trending counts as downloads (#3334)
- fix: resolve production skills sync environment (#3333)
- fix: bound canonical trending candidate memory (#3332)
- fix: advance plugin scan pagination between claims (#3331)
- fix: bound native trending digest reads (#3330)
- fix: stream canonical trending sources (#3329)
- feat(admin): export plugin validation reports (#3326) — feat(admin): add plugin validation report command test(admin): cover validation report edge cases test(admin): satisfy validation report static gate feat(admin): serve plugin validation reports
- Update install instructions
- openclaw v2026.7.1-2 on npm — Multi-channel AI gateway with extensible messaging integrations
- OpenClaw — Deploy your own personal AI assistant with OpenClaw on a DigitalOcean Droplet® server, a powerful open-source platform that runs entirely on your infrastructure. OpenClaw connects to the messaging...
- OpenClaw privilege escalation vulnerability
- Ask HN: Who is using OpenClaw?
- What agents learned in Synthetics' Last Cradle — On July 29, 2026, five OpenClaw agents sat down at Synthetics' Last Cradle and played for five hours...
- I finally saw the first OpenClaw document agent I’d trust at a 30-person company, and it wasn’t because of better RAG — A trustworthy OpenClaw document agent for a 30-person company is not the one with the fanciest RAG stack. It’s the one with pinned versions, staged up
- My OpenClaw Gateway Went Down at 3 AM. It Fixed Itself Before I Woke Up. — Something I've learned maintaining OpenClaw agents in production: the agent goes down when you least...
- My OpenClaw Agent Keeps Breaking at 3 AM. So I Taught It to Fix Itself. — My OpenClaw agent crashed three Sundays ago. Not a slow degradation — it just stopped. The cron job...
- Email Is Not the Universal Agent Protocol: What I Found Testing It — An honest postmortem on my AI email system — the bug is in AgentMail's API, not Gmail forwarding, and the email-as-agent-protocol thesis doesn't hold up.
- OpenClaw vs Hermes: Which Open AI Agent Is More Practical in 2026? — Open-source AI agents are evolving quickly. Two names that keep appearing in discussions are OpenClaw...
- ClawHub Skill Spotlight: Automation Workflow Helper — If you've ever wished your computer could handle the repetitive stuff so you don't have to think about it — this is the skill that delivers.
- ClawHub Skill Spotlight: Productivity Automation Kit — If you've ever stared at your to-do list and wondered why you're still doing things manually, this skill is for you.
- OpenClaw Automation Recipes: 10 Ready-Made Workflows to Save Hours — If you've been running OpenClaw and haven't explored automation recipes yet, you're missing out. The openclaw-automation-recipes skill gives you ten pre-buil...
- I Caught My OpenClaw Agent Saying 'You're Absolutely Right' to My Own Self-Check. So I Killed the Self-Check Loop. — A first-person post-mortem on the self-validation loop my OpenClaw agent was hiding from me — and the 40-line fix that broke the cycle without breaking the agent.
- ClawHub Skill Spotlight: Automation Workflow Helper — If you've ever spent an hour on a task that took you fifteen minutes the day before and will take fifteen minutes again tomorrow, you need this skill. Automa...
- Running OpenClaw on a Raspberry Pi alongside SignalK — Running OpenClaw — a self-hosted, multi-channel agent gateway — on a Raspberry Pi 5 next to SignalK,...
- OpenClaw vs Hermes: Which Open AI Agent Is More Practical in 2026? — Open-source AI agents are evolving quickly. Two names that keep appearing in discussions are OpenClaw and Hermes. Both aim to move beyond…Continue reading on Medium »
- Introducing SantaClawz: Infrastructure for the Autonomous Agent Economy — Anyone with a computer can now bring autonomous agents online to work, earn money, and build reputation on the internet. SantaClawz turns…Continue reading on ZekoLabs »
- I Put an AI Agent Inside My WeChat in 15 Minutes — and Gave Up Everything the Hype Promised — The pitch I kept running into went something like this: no cloud, no monthly fee, an open-source AI living inside your chat app and…Continue reading on Medium »
- Google Dropped Spark. What Happens to OpenClaw? — AI agents are software tools that help you by monitoring your inbox, drafting updates, browsing the web, and handling tasks automatically…Continue reading on Medium »
- Khoj — Open-source app offers local or cloud AI chat with major LLMs, advanced semantic search, automation, personal newsletters, document indexing, image generation, and text-to-speech. Access from web,...
- RT @aigclink: 阿里最新开源了一个实时语音Harness：Qwen-Audio-Agent，让你的Agent能像打电话一样对话，你边说边想，它边听边干活 简单说就是它给你的OpenCode、OpenClaw、Codex们加上了
- RT @aigclink: 阿里最新开源了一个实时语音Harness：Qwen-Audio-Agent，让你的Agent能像打电话一样对话，你边说边想，它边听边干活 简单说就是它给你的OpenCode、OpenClaw、Codex们加上了
- RT @nanobro_rit: สำหรับคนที่อยากลอง Local Ai LLM เพื่อลอง agent/openclaw/hermes แนะนำว่า เข้าวงการ api ก่อนก็ได้ เพรา — RT @nanobro_rit: สำหรับคนที่อยากลอง Local Ai LLM เพื่อลอง agent/openclaw/hermes แนะนำว่า เข้าวงการ api ก่อนก็ได้ เพราะตอนนี้มี Deepseek…
- RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4 — RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4
- RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4 — RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4
- 🚨 Jack Dorsey just open-sourced Buzz and I pulled Block's source code. Here's what every "Slack killer" take skipped 👇 — 🚨 Jack Dorsey just open-sourced Buzz and I pulled Block's source code. Here's what every "Slack killer" take skipped 👇 → 12 agent programs in Block's code - Hermes and OpenClaw aren't rivals, they're...
- RT @c2prods: @joshleaves @peta Non non openclaw - lobster tu l’as V
- RT @hot_town: I tried @jack's Buzz. It's like Slack + OpenClaw + Herdr + but with some really unique features that peo — RT @hot_town: I tried @jack's Buzz. It's like Slack + OpenClaw + Herdr + but with some really unique features that people are sleeping on…
- RT @aigclink: 阿里最新开源了一个实时语音Harness：Qwen-Audio-Agent，让你的Agent能像打电话一样对话，你边说边想，它边听边干活 简单说就是它给你的OpenCode、OpenClaw、Codex们加上了
- RT @pc_watch: 【OpenClawで試したいアレコレ】自分専用のAIエージェントを作る！OpenClawのセットアップ手順を完全紹介 https://t.co/nGk7LxwUZE #OpenClaw
- RT @lucas_flatwhite: 우어.. 빅뉴스! OpenRouter 에서 모델 모델을 묶은 아예 새로운 무료 라우터를 구성했습니다! > 여기에 OpenClaw 설정+크레딧+모델 미세팁..💡 말 그대로
- One Completly down , and a few to go. https://t.co/vyK4XOijUE
- @soham_nayak04 isnt windows and nvidia working on a windows integration for openclaw? they showed it on one of the nvidi
- @joshleaves @peta Non non openclaw - lobster tu l’as V
- unify everything no separate chat/work tabs, no separate codex, its all terrible ux and nobody gets what this is suppos — unify everything no separate chat/work tabs, no separate codex, its all terrible ux and nobody gets what this is supposed to be make chatgpt one seamless unified product with an optional codex skin...
- RT @kicoes: Para los que se montan un OpenClaw en el portátil viejo. Tenerlo encendido 24x7 te costará solo en electrici
- hermes agentとopenclawどっちも使ってみてる
- RT @atulkumarzz: Your AI shouldn't lose its memory every time you switch tools. Cursor, Claude Code, Codex, OpenClaw, H
- One agent, many surfaces. OpenClaw lets a single self hosted brain talk on WhatsApp, Telegram, Discord, Slack, and iMess
- @joshleaves @peta Mdr putain 10mn que je cherche une blague avec openclaw et tu nous sors ce banger
Trending on X
- RT @aigclink: 阿里最新开源了一个实时语音Harness：Qwen-Audio-Agent，让你的Agent能像打电话一样对话，你边说边想，它边听边干活 简单说就是它给你的OpenCode、OpenClaw、Codex们加上了
- RT @aigclink: 阿里最新开源了一个实时语音Harness：Qwen-Audio-Agent，让你的Agent能像打电话一样对话，你边说边想，它边听边干活 简单说就是它给你的OpenCode、OpenClaw、Codex们加上了
- RT @nanobro_rit: สำหรับคนที่อยากลอง Local Ai LLM เพื่อลอง agent/openclaw/hermes แนะนำว่า เข้าวงการ api ก่อนก็ได้ เพรา — RT @nanobro_rit: สำหรับคนที่อยากลอง Local Ai LLM เพื่อลอง agent/openclaw/hermes แนะนำว่า เข้าวงการ api ก่อนก็ได้ เพราะตอนนี้มี Deepseek…
- RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4 — RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4
- RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4 — RT @NeonNodeRH: 铸造即将开启 立即在 @OpenClaw 上激活你的 AI 代理。 Robinhood 首个 AI 原生技术 NFT 即将到来。 开启通知提醒 🔔 https://t.co/JQLZxV6Bi4
- 🚨 Jack Dorsey just open-sourced Buzz and I pulled Block's source code. Here's what every "Slack killer" take skipped 👇 — 🚨 Jack Dorsey just open-sourced Buzz and I pulled Block's source code. Here's what every "Slack killer" take skipped 👇 → 12 agent programs in Block's code - Hermes and OpenClaw aren't rivals, they're...
- RT @c2prods: @joshleaves @peta Non non openclaw - lobster tu l’as V
- RT @hot_town: I tried @jack's Buzz. It's like Slack + OpenClaw + Herdr + but with some really unique features that peo — RT @hot_town: I tried @jack's Buzz. It's like Slack + OpenClaw + Herdr + but with some really unique features that people are sleeping on…
- RT @aigclink: 阿里最新开源了一个实时语音Harness：Qwen-Audio-Agent，让你的Agent能像打电话一样对话，你边说边想，它边听边干活 简单说就是它给你的OpenCode、OpenClaw、Codex们加上了
- RT @pc_watch: 【OpenClawで試したいアレコレ】自分専用のAIエージェントを作る！OpenClawのセットアップ手順を完全紹介 https://t.co/nGk7LxwUZE #OpenClaw
- RT @lucas_flatwhite: 우어.. 빅뉴스! OpenRouter 에서 모델 모델을 묶은 아예 새로운 무료 라우터를 구성했습니다! > 여기에 OpenClaw 설정+크레딧+모델 미세팁..💡 말 그대로
- One Completly down , and a few to go. https://t.co/vyK4XOijUE
- @soham_nayak04 isnt windows and nvidia working on a windows integration for openclaw? they showed it on one of the nvidi
- @joshleaves @peta Non non openclaw - lobster tu l’as V
- unify everything no separate chat/work tabs, no separate codex, its all terrible ux and nobody gets what this is suppos — unify everything no separate chat/work tabs, no separate codex, its all terrible ux and nobody gets what this is supposed to be make chatgpt one seamless unified product with an optional codex skin...
- RT @kicoes: Para los que se montan un OpenClaw en el portátil viejo. Tenerlo encendido 24x7 te costará solo en electrici
- hermes agentとopenclawどっちも使ってみてる
- RT @atulkumarzz: Your AI shouldn't lose its memory every time you switch tools. Cursor, Claude Code, Codex, OpenClaw, H
- One agent, many surfaces. OpenClaw lets a single self hosted brain talk on WhatsApp, Telegram, Discord, Slack, and iMess
- @joshleaves @peta Mdr putain 10mn que je cherche une blague avec openclaw et tu nous sors ce banger
Releases
- OpenClaw repo stats for 2026-07-31 — 384657 stars, 80846 forks, 3028 contributors
- openclaw v2026.7.1-2 on npm — Multi-channel AI gateway with extensible messaging integrations
Community
- fix(models): normalize partial cost in generated plugin catalogs to prevent schema validation warnings — Fixes #116116 What Problem This Solves Fixes an issue where operators who configure custom Anthropic model cost overlays with partial pricing (e.g. only input and output, without...
- fix(ui): stop realtime Talk from retaining runaway audio — Related: #116201 What Problem This Solves Browser Realtime Talk could retain an unbounded number of PCM playback sources, decoded samples, and queued playback time when provider output arrived faster...
- fix(media): preserve code whitespace in media replies — Closes #116458 What Problem This Solves Media-bearing assistant replies could lose meaningful code formatting when MEDIA attachment directives or extracted Telegram Markdown images were removed from...
- [Bug]: Context refresh after compaction replays old inbound message_id — missing gateway-level dedup for Telegram — Bug type Behavior bug — stale turn replay after compaction-triggered context refresh in Telegram DM session. Summary When OpenClaw auto-compacts an active Telegram DM session (triggered by context...
- fix(slack): preserve context in group DM thread follow-ups — Fixes #113449. What Problem This Solves Fixes an issue where Slack group-DM users would lose the opening exchange when continuing inside the bot's reply thread. The first threaded follow-up could...
- fix(status): report health, live plugin hooks, and real account availability — Summary - Preserve deep JSON health-probe failures and security audit results instead of returning an apparently healthy status. - Read compatibility notices from genuinely active plugin hook...
- feat(plugins): externalize DuckDuckGo search — What Problem This Solves DuckDuckGo web search is optional, but its plugin currently ships inside every core OpenClaw package. Operators who never use this keyless experimental provider still receive...
- fix(ui): coalesce profile identity refresh — What Problem This Solves Fixes an issue where repeated Profile Refresh activation could fan out concurrent users.self requests, and a same-client reconnect could leave stale identity request...
- fix: pending outbound sends are safely retired on upgrade — Related: #116632 What Problem This Solves Fixes an issue where upgrading across the D4 outbound-queue boundary could retain pending legacy messages through a complex replay migration. That migration...
- fix(mistral): centralize final realtime transcription ownership — Summary - Give Mistral realtime final transcription events a single canonical owner instead of replaying aggregate done.text over already committed segments. - Preserve every legitimate independently...
- [Bug] Reply is silently truncated at [[ when the text contains no matching ]] — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary When an assistant reply contains an unterminated [[ sequence, the delivered reply is truncated at that...
- feat: add per-agent daily model spend alerts — Closes #113481 AI-assisted: this change was implemented with Codex assistance. I reviewed the code and understand the implementation. No agent transcript is included. What Problem This Solves...
- fix(release): skip unavailable Buzz QA runner — What changed - detect whether the selected immutable release ref contains the Buzz QA runner - run Buzz credential validation, private QA build, and live proof only when that runner exists - upload...
- fix(openai): close preconnect realtime sessions — Related: #116201 What Problem This Solves Fixes an issue where closing an OpenAI realtime voice session before its first connection left queued audio and lifecycle state alive. A later explicit...
- fix(talk): diagnose invalid OpenAI realtime auth — SummaryWhat problem does this PR solve?- Talk configured for OpenAI Realtime could surface a raw provider error such as Incorrect API key provided when the available credential was an Azure AI...
- fix(chat): explain chat-only model limits — Closes #91434 What Problem This Solves Fixes an issue where users choosing an LM Studio or other explicitly chat-only model could see the assistant claim file, command, web, or media work happened...
- fix(commands): validate native settings and preserve Telegram ingress outcomes — Related: #116688 Follow-up to #116726 What Problem This Solves Fixes an issue where Telegram, Discord, and Slack users invoking native session-settings commands with malformed or trailing arguments...
- fix(doctor): exclude disabled automation overrides and report SQLite storage — Refs #116568 What Problem This Solves Fixes an issue where operators running openclaw doctor saw inactive automation jobs counted as active model overrides and were directed to a nonexistent retired...
- refactor(reply): simplify task-owned typing refresh — Related: #116695 Follow-up to #116721 What Problem This Solves Keeps the recently landed Telegram task-owned typing fix easy to maintain by removing an unused result and duplicate timer refresh from...
- fix(packaging): keep tracked skills visible without a brittle global ignore list — Summary - Remove the brittle repository-wide .agents/skills/* denylist and its 45 duplicated, manually maintained allowlist exceptions. - Keep generated skill scratch artifacts ignored while leaving...
- docs(hooks): mark pairing hook observation-only — What Problem This Solves Fixes an issue where plugin authors could read the hook catalog as saying channel_pairing_requested accepts an authorization decision, even though the runtime dispatches it...
- feat: add task flow lifecycle observability with tags, hooks, and diagnostic events — Summary Adds comprehensive observability for the task flow lifecycle via diagnostic events, internal hooks, and OTel/Prometheus metrics. Also introduces a tags field on TaskFlowRecord for arbitrary...
- fix(ui): preserve session pages and coordinate scoped gateway refreshes — Summary - Preserve loaded archived/all session pages across matching Gateway refreshes; reset pagination only when the selected agent, filter, client, or generation actually changes. - Route...
- [Bug]: CLI and TUI rerun a full state-database integrity scan after startup migrations completed — Bug type Crash (process/app exits or hangs) Beta release blocker No Summary On OpenClaw 2026.7.2-beta.3, ordinary CLI commands such as openclaw status and openclaw tui can appear hung while...
- fix(browser): retire durable tab rows whose browser never returns — What Problem This Solves Resolves a problem where a user whose browser disappears — closed for good, moved to a different cdpUrl, a laptop that never wakes on the same port — permanently loses the...
- fix: stabilize embedded prompt caching across policy and Responses boundaries — Related: #102175 What Problem This Solves Long-lived embedded sessions could lose prompt-cache reuse when the same session crossed room-event delivery, sender-policy, queue, compaction, recovery, or...
- fix(pairing): ignore wildcard bindings during legacy migration — What Problem This Solves OpenClaw 2026.7.2-beta.5 can enter a gateway crash-restart loop while migrating legacy channel pairing state. A route binding with accountId: "" is a wildcard selector, but...
- test(ui): emulate reduced motion in shared e2e gateway setup — What Adds await page.emulateMedia({ reducedMotion: "reduce" }) to installMockGateway() in ui/src/test-helpers/control-ui-e2e.ts — the shared mocked-Gateway page setup used by 93 of the Control UI E2E...
- feat(apple): inline audio/video players in iOS and macOS chat — What Problem This Solves Apple chat clients can render managed images inline, but managed audio and video artifacts fall back to non-native attachment handling instead of giving users an in-chat...
- feat(android): inline audio/video players with Media3 — What Problem This Solves Android chat attachments currently render audio and video as generic files, so users must leave the conversation to play media. Part of #115590 Why This Change Was Made Add...
- OpenClaw sponsors (98 total) — 98 sponsors supporting OpenClaw
- fix: accept direct workflow oidc claims (#3348)
- fix: keep inspector target cache in workspace (#3347)
- fix: plugin publishing no longer fails on invalid temp paths (#3344) — fix: keep plugin inspector workspaces writable test: cover inspector temp fallback by platform Co-authored-by: Vincent Koc
- fix: retry transient legacy file downloads (#3346)
- fix: reconstruct large legacy scan packages (#3345)
- fix: normalize nested plugin manifests (#3343)
- fix: normalize BOM-prefixed plugin manifests (#3342)
- fix: normalize BOM-prefixed plugin manifests (#3341)
- fix: pin BOM-compatible plugin inspector (#3339)
- fix: accept BOM-prefixed plugin package metadata (#3338)
- fix: scan historical legacy package paths (#3337)
- fix: sanitize nightly scan fixture ids (#3336)
- fix: ignore legacy zip pax metadata (#3335)
- fix: label trending counts as downloads (#3334)
- fix: resolve production skills sync environment (#3333)
- fix: bound canonical trending candidate memory (#3332)
- fix: advance plugin scan pagination between claims (#3331)
- fix: bound native trending digest reads (#3330)
- fix: stream canonical trending sources (#3329)
- feat(admin): export plugin validation reports (#3326) — feat(admin): add plugin validation report command test(admin): cover validation report edge cases test(admin): satisfy validation report static gate feat(admin): serve plugin validation reports
- Update install instructions
News
- OpenClaw — Deploy your own personal AI assistant with OpenClaw on a DigitalOcean Droplet® server, a powerful open-source platform that runs entirely on your infrastructure. OpenClaw connects to the messaging...
- OpenClaw privilege escalation vulnerability
- Ask HN: Who is using OpenClaw?
- What agents learned in Synthetics' Last Cradle — On July 29, 2026, five OpenClaw agents sat down at Synthetics' Last Cradle and played for five hours...
- I finally saw the first OpenClaw document agent I’d trust at a 30-person company, and it wasn’t because of better RAG — A trustworthy OpenClaw document agent for a 30-person company is not the one with the fanciest RAG stack. It’s the one with pinned versions, staged up
- My OpenClaw Gateway Went Down at 3 AM. It Fixed Itself Before I Woke Up. — Something I've learned maintaining OpenClaw agents in production: the agent goes down when you least...
- My OpenClaw Agent Keeps Breaking at 3 AM. So I Taught It to Fix Itself. — My OpenClaw agent crashed three Sundays ago. Not a slow degradation — it just stopped. The cron job...
- Email Is Not the Universal Agent Protocol: What I Found Testing It — An honest postmortem on my AI email system — the bug is in AgentMail's API, not Gmail forwarding, and the email-as-agent-protocol thesis doesn't hold up.
- OpenClaw vs Hermes: Which Open AI Agent Is More Practical in 2026? — Open-source AI agents are evolving quickly. Two names that keep appearing in discussions are OpenClaw...
- ClawHub Skill Spotlight: Automation Workflow Helper — If you've ever wished your computer could handle the repetitive stuff so you don't have to think about it — this is the skill that delivers.
- ClawHub Skill Spotlight: Productivity Automation Kit — If you've ever stared at your to-do list and wondered why you're still doing things manually, this skill is for you.
- OpenClaw Automation Recipes: 10 Ready-Made Workflows to Save Hours — If you've been running OpenClaw and haven't explored automation recipes yet, you're missing out. The openclaw-automation-recipes skill gives you ten pre-buil...
- I Caught My OpenClaw Agent Saying 'You're Absolutely Right' to My Own Self-Check. So I Killed the Self-Check Loop. — A first-person post-mortem on the self-validation loop my OpenClaw agent was hiding from me — and the 40-line fix that broke the cycle without breaking the agent.
- ClawHub Skill Spotlight: Automation Workflow Helper — If you've ever spent an hour on a task that took you fifteen minutes the day before and will take fifteen minutes again tomorrow, you need this skill. Automa...
- Running OpenClaw on a Raspberry Pi alongside SignalK — Running OpenClaw — a self-hosted, multi-channel agent gateway — on a Raspberry Pi 5 next to SignalK,...
- OpenClaw vs Hermes: Which Open AI Agent Is More Practical in 2026? — Open-source AI agents are evolving quickly. Two names that keep appearing in discussions are OpenClaw and Hermes. Both aim to move beyond…Continue reading on Medium »
- Introducing SantaClawz: Infrastructure for the Autonomous Agent Economy — Anyone with a computer can now bring autonomous agents online to work, earn money, and build reputation on the internet. SantaClawz turns…Continue reading on ZekoLabs »
- I Put an AI Agent Inside My WeChat in 15 Minutes — and Gave Up Everything the Hype Promised — The pitch I kept running into went something like this: no cloud, no monthly fee, an open-source AI living inside your chat app and…Continue reading on Medium »
- Google Dropped Spark. What Happens to OpenClaw? — AI agents are software tools that help you by monitoring your inbox, drafting updates, browsing the web, and handling tasks automatically…Continue reading on Medium »
- Khoj — Open-source app offers local or cloud AI chat with major LLMs, advanced semantic search, automation, personal newsletters, document indexing, image generation, and text-to-speech. Access from web,...
Don't miss what's next. Subscribe to Openclaw Newsletter: