OpenClaw Newsletter - 2026-05-02
OpenClaw Newsletter
Saturday, May 2, 2026Top Stories
- OpenClaw 2026.4.29-beta.2 — 2026.4.29 Highlights - Messaging and automation get active-run steering by default, visible-reply enforcement, spawned subagent routing metadata, and opt-in follow-up commitments for...
- fix(control-ui): create sessions for typed /new — Summary - Route typed Control UI /new through the same dashboard session create-and-switch flow as the New Chat button. - Keep typed /reset on the explicit in-place gateway reset path. - Document the...
- fix(gateway): render codex inline media once — Summary - Problem: Codex-harness replies that emitted local MEDIA: image directives could fail to become trusted outgoing media, leaving inline images text-only or unavailable to WebChat/Control UI....
- feat(browser): add --headed flag to browser start for runtime headed handoff — Fixes #75879. What changes Exposes the existing POST /start?headless=false route via three public surfaces that previously only offered the headless direction: 1. CLI (--headed flag) Symmetric...
- fix(discord): harden Carbon parity — Summary - add Carbon-style Discord REST request lanes with stale background drop while preserving the existing global queue cap - harden gateway send handling with payload-size validation,...
- fix(heartbeat): centralized cooldown gate prevents exec-event runaway — Summary - Centralized cooldown gate prevents exec-event runaway loop in heartbeat dispatcher (#64016) - Discriminated-union WakeReason replaces raw string compare for reason classification -...
- perf(agents): memoize resolveTranscriptPolicy per config — Summary Memoize resolveTranscriptPolicy per OpenClawConfig object identity. Profiling on a stable warm gateway shows it synchronously costs ~0.9 sec per embedded turn even when nothing about the...
- fix(discord): document mention formatting guidance — Summary - Carry forward the Discord mention-formatting prompt guidance from source PR #74506. - Document user, channel, and role mention syntax in the Discord channel docs. - Add a focused Discord...
- feat: add scoped mention pattern policy — Summary - add a shared scoped mention-pattern policy for global, agent, and provider/channel overrides - apply configured mention-pattern gating across channel runtimes while preserving native...
- Fix doctor SecretRef false positive and normalize task timestamps — Summary - resolve SecretRef-managed gateway.auth.token before doctor decides whether token auth is missing - avoid false-positive doctor guidance when the SecretRef resolves successfully - normalize...
- perf(agents/pi-embedded): memoize resolvePreparedExtraParams per config — Summary Memoize resolvePreparedExtraParams per OpenClawConfig object identity. Profiling on a stable warm gateway shows it synchronously costs ~1.9 sec per embedded turn even when nothing about the...
- Squash test
- fix(ui): surface slash command dispatch failures — Summary - Surface a visible inline message when a local Control UI slash command cannot run because the Gateway client is unavailable. - Catch unexpected local slash-command dispatch failures and...
- Regression: 2026.4.27 artifacts lack public/internal assistant audience boundary for visible delivery — Summary The 2026.4.27 npm artifacts checked on 2026-04-29 appear to preserve some commentary / final_answer phase separation, but do not expose or enforce an explicit public/internal assistant...
- Embedded-run "auth" stage takes 10–15s synchronously regardless of model auth profile state
- [WIP] test: add session linkage verification for Slack roomish channels with replyToMode=all — Summary - Problem: Slack roomish channels ignore replyToMode="all", causing top-level messages to share a single session. - Why it matters: This leads to race conditions where responses to different...
- fix(kimi): switch to openai-completions endpoint for image support — Summary - Problem: The bundled kimi provider sent requests via Anthropic Messages format (/messages path), which Kimi's backend rejects with 403 Forbidden when the body contains image content. - Why...
- [Bug]: System very buggy, unstable and slow since at least v2026.4.24 — Bug type Regression (worked before, now fails) Beta release blocker No Summary I don't know if it was before 2026.04.24, that's the point where I have noticed it. I think a lot of the tickets...
- fix(plugins): drop duplicate official flag in plugins search output — Summary Removes destined-to-be-duplicate word "official" from plugin search results AI Generated description: - openclaw plugins search was rendering the isOfficial boolean as the literal string...
- fix(control-ui): measure and decouple slow refreshes — Summary - Add Control UI timing events for tab visible render, refresh scopes, secondary refresh groups, and attributed gateway RPCs. - Decouple primary tab feedback from expensive secondary...
- [Bug]: 4.29 dispatch prep stages take ~73s of synchronous CPU work, blocking event loop
- [Bug] Dreaming does not process the primary agent workspace — When memory-core dreaming runs at 03:00 UTC, the primary agent workspace (/root/.openclaw/workspace) is not included in the dreaming workspace enumeration. Only the 4 AGI sub-agents (agi-ceo,...
- fix: hot reload plugin management changes — Summary - add Gateway plugin hot reload for install/enable/disable/list/search management paths while keeping source-changing installs restart-backed - stop affected channel plugins before registry...
- fix(gateway): bound websocket auth after valid connect — Repair contributor PR #49751 in place. Plan: - Rebase the existing branch onto current main 381c2e1d1a3bbc5c41383e6d66b094bbdf3a6bf0. - Keep clearing the preauth timer only after a valid connect...
- [Bug]: Multiple regressions in 2026.4.8 - Gateway high CPU, CLI timeout, Discord voice broken, Messenger bot broken
- fix(tts): honor short explicit tagged speech text — Summary - honor explicitly tagged [[tts:text]]...[[/tts:text]] content even when it is short - keep the existing short-text guard for untagged auto-TTS replies - add regression coverage for both...
- [Bug]: TTS directives ( + [[tts:text]]) silently dropped on claude-cli backend; no voice delivere — Bug type Regression (worked before, now fails) Beta release blocker No Summary On openclaw 2026.4.26 with the claude-cli backend, model output containing and [[tts:text]]…[[/tts:text]] is parsed but...
- iMessage monitor: no per-turn watchdog for tool activity during agent silence — Symptom When an agent is silent past a threshold but is still running tools, the iMessage monitor sends nothing — the user has no signal that work is in progress. On longer turns this looks...
- Elevated exec still flaps after PR #74666 (v2026.4.29) — Summary Elevated exec permissions drop after the first approved command completes in the same assistant turn. PR #74666 ("preserve turnSourceChannel as messageProvider in approval followup runs") was...
- [Bug]: 2026.4.29 Web UI chat bubbles then disappears; direct OpenAI API works — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Hi OpenClaw Support, I’m having an issue where OpenClaw loads normally, but the Web UI chat does not...
- Task registry maintenance can burn CPU quadratically with many tasks — Summary OpenClaw 2026.4.29 can spend sustained CPU in task registry maintenance when the task registry contains many records. What I observed While validating an owned 2026.4.29 build on macOS, the...
- OpenClaw repo stats for 2026-05-02 — 367317 stars, 75522 forks, 2048 contributors
- OpenClaw sponsors (123 total) — 123 sponsors supporting OpenClaw
- fix: add package identity repair admin
- fix: allow JSON Schema manifests in package publish
- fix: keep beta plugin packages off latest
- fix: add package owner transfer repair
- feat: reserve OpenClaw plugin package names
- Update securityPrompt.ts
- Update securityPrompt.ts
- Update securityPrompt.ts
- fix: rely on JSON artifact neutralization
- Update securityPrompt.ts
- fix: wrap ClawScan skill artifacts in prompt boundary
- ci: expand pr validation coverage — Split PR validation into explicit static, unit, package, type/build, HTTP e2e, and browser-smoke gates. Add local ci:* scripts and document the required status checks.
- fix: keep probe helper types private
- openclaw-cli 2026.4.29 on Homebrew — Your own personal AI assistant
- OpenClaw on DigitalOcean — 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...
- Google restricting Google AI Pro/Ultra subscribers for using OpenClaw
- OpenClaw is a security nightmare dressed up as a daydream
- OpenClaw is changing my life
- OpenClaw’s memory is unreliable, and you don’t know when it will break
- This is the most capable, highest level of sophistication agentic framework in the world right now. — I open sourced it yesterday. What makes that sentence even stranger is this. The agent itself built and published its own repository on…Continue reading on Medium »
- Testing Documentation Like Software in an Air-Gapped Environment — Most documentation does not fail because the writer was careless.Continue reading on Medium »
- MCP Is Not an API Layer. It’s an Agent Economy. — The industry keeps describing Model Context Protocol (MCP) as “agents calling tools.” That framing is convenient but dangerously…Continue reading on Medium »
- Running OpenClaw in a Docker Container on Ubuntu 22.04 — If you desire, you can follow the documentation that OpenClaw has proposed themself. You can find this documentation here.Continue reading on Medium »
- Hermes vs. OpenClaw: Which Open-Source AI Agent Really Wins? — 127K stars in 9 weeks. Here’s why Hermes is the fastest-growing AI agent on GitHub — and how it differs from the 345K-star veteran…Continue reading on Medium »
- OpenClaw, AI AGENT로 24/7 목표를 향해 스스로 움직이는 공장 — “ERP부터 MES까지를 AI가 관통하는 자율 생산 시스템”Continue reading on Medium »
- Claude Is Down Again? Here’s The Solution — Claude has been going down a lot lately. So much that it only has one 9 of uptime.Continue reading on Medium »
- My OpenClaw agent didn’t get dumber — I just gave it 50 skills and hoped for the best — A lot of “my OpenClaw agent got worse” stories start the same way: somebody stuffed 50 overlapping skills into one agent and called it…Continue reading on Medium »
- The Uniformed Guard Problem: Why AI Agent Sandboxes Need Identity, Not Just Policy — NemoClaw is NVIDIA’s reference stack for running OpenClaw agents safely. It wraps the agent in an OpenShell sandbox with a deny-by-default…Continue reading on Highflame »
- Your Hermes Agent Has No Performance Review. Here’s How to Fix That — Everyone is excited about the self-learning loop. Nobody is talking about what happens when it learns the wrong thing. Rollback strategies…Continue reading on Medium »
- How OpenClaw Agents Bypass Bot Detection (And How to Stop Them) (4 minute read) — May 01|InfosecHow OpenClaw Agents Bypass Bot Detection (And How to Stop Them) (4 minute read)
- remember people, if you wanted a openclaw agent for yourself, you can deploy one in minutes through https://t.co/uPGfCRW
- @wanerfu Built ClawHost so you can deploy OpenClaw with one click and get full access to all its features including scra
- OpenClaw is basically that one chaotic genius friend: somehow ships useful stuff at 2am and leaves you wondering if you’
- @BLISEARTH 😃😄😂 I sent this using openClaw It was supposed to be 2k 😂😂😂
- I'm just getting started with @openclaw . I setup a bot on my discord with Gemma 4 e4b LLM running locally on my laptop.
- OpenClaw：能力强，但每个细节都得我盯着。 Hermes Agent：更像我的延伸大脑... 一句话：更丝滑，更省心。 #AI工具 #效率提升 https://t.co/fsVVEV1qWJ — OpenClaw：能力强，但每个细节都得我盯着。 Hermes Agent：更像我的延伸大脑... 一句话：更丝滑，更省心。 #AI工具 #效率提升 https://t.co/fsVVEV1qWJ
- 사이버-레닌은 OpenClaw, Hermes 같은 범용 오픈소스 하네스(다운받은 적도 없음) 대신 연초부터 자체 하네스 만들어서 돌리는 중인데... 그래서 버그도 내가 잡고 개선점도 내가 찾아서 적용해야하는..
- @0xbl33p @fxnction @openclaw Support the token...it will create headlines brother
- 次にスマホのroot化しようとしたけどopenclawはそんなこと出来ませんて出た。クソ、また調べ直しか。もう疲れたのでまた明日だな。 https://t.co/xIk5dC9kRX
- プロンプト集100個以上！？ ChatGPT OpenAI 初心者向けの画像生成AI プロンプト集 ChatGPT Images 2.0×Codex imagegenとClaudeCodeでAI画像生成を自動化まで徹底解説 【完全保存版】C — プロンプト集100個以上！？ ChatGPT OpenAI 初心者向けの画像生成AI プロンプト集 ChatGPT Images 2.0×Codex imagegenとClaudeCodeでAI画像生成を自動化まで徹底解説 【完全保存版】Claude Design GPT-5.5 Claude Code OpenClaw Antigravity Codex Obsibian Cursor...
- Why is no one talking about n8n or openclaw anymore
- @designertom @openclaw Remember, if you wanted your own openclaw agent, you can deploy one in minutes through https://t.
- 普通人也能用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/9DNILYE13O — 普通人也能用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/9DNILYE13O
- Time for another go at openclaw — let’s see what breaks this time. https://t.co/lQZxZ9fuAj
- @tomfgoodwin Yep, already happening. Maybe you need to get your openclaw to do some research for you. — @tomfgoodwin Yep, already happening. Maybe you need to get your openclaw to do some research for you.
- @sama Someone give me link to openclaw @grok
- 入坑 #openclaw #hermes #ai https://t.co/Qwh1d6xhvX — 入坑 #openclaw #hermes #ai https://t.co/Qwh1d6xhvX
- 65NroJtTcffbb2pJRNZMLAdAyakKv4nj9eAKZZp8pump openclaw flagged $ADS early. price is moving exactly as predicted Was M
- 为什么要用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/SLi21kcNsS — 为什么要用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/SLi21kcNsS
- @bugtrader69 @BRamOL yeah, but the point was that all you have to do is pay a monthly subscription and you can just conn
Trending on X
- remember people, if you wanted a openclaw agent for yourself, you can deploy one in minutes through https://t.co/uPGfCRW
- @wanerfu Built ClawHost so you can deploy OpenClaw with one click and get full access to all its features including scra
- OpenClaw is basically that one chaotic genius friend: somehow ships useful stuff at 2am and leaves you wondering if you’
- @BLISEARTH 😃😄😂 I sent this using openClaw It was supposed to be 2k 😂😂😂
- I'm just getting started with @openclaw . I setup a bot on my discord with Gemma 4 e4b LLM running locally on my laptop.
- OpenClaw：能力强，但每个细节都得我盯着。 Hermes Agent：更像我的延伸大脑... 一句话：更丝滑，更省心。 #AI工具 #效率提升 https://t.co/fsVVEV1qWJ — OpenClaw：能力强，但每个细节都得我盯着。 Hermes Agent：更像我的延伸大脑... 一句话：更丝滑，更省心。 #AI工具 #效率提升 https://t.co/fsVVEV1qWJ
- 사이버-레닌은 OpenClaw, Hermes 같은 범용 오픈소스 하네스(다운받은 적도 없음) 대신 연초부터 자체 하네스 만들어서 돌리는 중인데... 그래서 버그도 내가 잡고 개선점도 내가 찾아서 적용해야하는..
- @0xbl33p @fxnction @openclaw Support the token...it will create headlines brother
- 次にスマホのroot化しようとしたけどopenclawはそんなこと出来ませんて出た。クソ、また調べ直しか。もう疲れたのでまた明日だな。 https://t.co/xIk5dC9kRX
- プロンプト集100個以上！？ ChatGPT OpenAI 初心者向けの画像生成AI プロンプト集 ChatGPT Images 2.0×Codex imagegenとClaudeCodeでAI画像生成を自動化まで徹底解説 【完全保存版】C — プロンプト集100個以上！？ ChatGPT OpenAI 初心者向けの画像生成AI プロンプト集 ChatGPT Images 2.0×Codex imagegenとClaudeCodeでAI画像生成を自動化まで徹底解説 【完全保存版】Claude Design GPT-5.5 Claude Code OpenClaw Antigravity Codex Obsibian Cursor...
- Why is no one talking about n8n or openclaw anymore
- @designertom @openclaw Remember, if you wanted your own openclaw agent, you can deploy one in minutes through https://t.
- 普通人也能用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/9DNILYE13O — 普通人也能用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/9DNILYE13O
- Time for another go at openclaw — let’s see what breaks this time. https://t.co/lQZxZ9fuAj
- @tomfgoodwin Yep, already happening. Maybe you need to get your openclaw to do some research for you. — @tomfgoodwin Yep, already happening. Maybe you need to get your openclaw to do some research for you.
- @sama Someone give me link to openclaw @grok
- 入坑 #openclaw #hermes #ai https://t.co/Qwh1d6xhvX — 入坑 #openclaw #hermes #ai https://t.co/Qwh1d6xhvX
- 65NroJtTcffbb2pJRNZMLAdAyakKv4nj9eAKZZp8pump openclaw flagged $ADS early. price is moving exactly as predicted Was M
- 为什么要用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/SLi21kcNsS — 为什么要用AI员工 AI 员工不是聊天机器人，而是能接工具、记忆上下文、执行任务的数字员工。 #OpenClaw #AI员工 #Agent #AI自动化 https://t.co/SLi21kcNsS
- @bugtrader69 @BRamOL yeah, but the point was that all you have to do is pay a monthly subscription and you can just conn
Releases
- OpenClaw 2026.4.29-beta.2 — 2026.4.29 Highlights - Messaging and automation get active-run steering by default, visible-reply enforcement, spawned subagent routing metadata, and opt-in follow-up commitments for...
- OpenClaw repo stats for 2026-05-02 — 367317 stars, 75522 forks, 2048 contributors
- openclaw-cli 2026.4.29 on Homebrew — Your own personal AI assistant
Community
- fix(control-ui): create sessions for typed /new — Summary - Route typed Control UI /new through the same dashboard session create-and-switch flow as the New Chat button. - Keep typed /reset on the explicit in-place gateway reset path. - Document the...
- fix(gateway): render codex inline media once — Summary - Problem: Codex-harness replies that emitted local MEDIA: image directives could fail to become trusted outgoing media, leaving inline images text-only or unavailable to WebChat/Control UI....
- feat(browser): add --headed flag to browser start for runtime headed handoff — Fixes #75879. What changes Exposes the existing POST /start?headless=false route via three public surfaces that previously only offered the headless direction: 1. CLI (--headed flag) Symmetric...
- fix(discord): harden Carbon parity — Summary - add Carbon-style Discord REST request lanes with stale background drop while preserving the existing global queue cap - harden gateway send handling with payload-size validation,...
- fix(heartbeat): centralized cooldown gate prevents exec-event runaway — Summary - Centralized cooldown gate prevents exec-event runaway loop in heartbeat dispatcher (#64016) - Discriminated-union WakeReason replaces raw string compare for reason classification -...
- perf(agents): memoize resolveTranscriptPolicy per config — Summary Memoize resolveTranscriptPolicy per OpenClawConfig object identity. Profiling on a stable warm gateway shows it synchronously costs ~0.9 sec per embedded turn even when nothing about the...
- fix(discord): document mention formatting guidance — Summary - Carry forward the Discord mention-formatting prompt guidance from source PR #74506. - Document user, channel, and role mention syntax in the Discord channel docs. - Add a focused Discord...
- feat: add scoped mention pattern policy — Summary - add a shared scoped mention-pattern policy for global, agent, and provider/channel overrides - apply configured mention-pattern gating across channel runtimes while preserving native...
- Fix doctor SecretRef false positive and normalize task timestamps — Summary - resolve SecretRef-managed gateway.auth.token before doctor decides whether token auth is missing - avoid false-positive doctor guidance when the SecretRef resolves successfully - normalize...
- perf(agents/pi-embedded): memoize resolvePreparedExtraParams per config — Summary Memoize resolvePreparedExtraParams per OpenClawConfig object identity. Profiling on a stable warm gateway shows it synchronously costs ~1.9 sec per embedded turn even when nothing about the...
- Squash test
- fix(ui): surface slash command dispatch failures — Summary - Surface a visible inline message when a local Control UI slash command cannot run because the Gateway client is unavailable. - Catch unexpected local slash-command dispatch failures and...
- Regression: 2026.4.27 artifacts lack public/internal assistant audience boundary for visible delivery — Summary The 2026.4.27 npm artifacts checked on 2026-04-29 appear to preserve some commentary / final_answer phase separation, but do not expose or enforce an explicit public/internal assistant...
- Embedded-run "auth" stage takes 10–15s synchronously regardless of model auth profile state
- [WIP] test: add session linkage verification for Slack roomish channels with replyToMode=all — Summary - Problem: Slack roomish channels ignore replyToMode="all", causing top-level messages to share a single session. - Why it matters: This leads to race conditions where responses to different...
- fix(kimi): switch to openai-completions endpoint for image support — Summary - Problem: The bundled kimi provider sent requests via Anthropic Messages format (/messages path), which Kimi's backend rejects with 403 Forbidden when the body contains image content. - Why...
- [Bug]: System very buggy, unstable and slow since at least v2026.4.24 — Bug type Regression (worked before, now fails) Beta release blocker No Summary I don't know if it was before 2026.04.24, that's the point where I have noticed it. I think a lot of the tickets...
- fix(plugins): drop duplicate official flag in plugins search output — Summary Removes destined-to-be-duplicate word "official" from plugin search results AI Generated description: - openclaw plugins search was rendering the isOfficial boolean as the literal string...
- fix(control-ui): measure and decouple slow refreshes — Summary - Add Control UI timing events for tab visible render, refresh scopes, secondary refresh groups, and attributed gateway RPCs. - Decouple primary tab feedback from expensive secondary...
- [Bug]: 4.29 dispatch prep stages take ~73s of synchronous CPU work, blocking event loop
- [Bug] Dreaming does not process the primary agent workspace — When memory-core dreaming runs at 03:00 UTC, the primary agent workspace (/root/.openclaw/workspace) is not included in the dreaming workspace enumeration. Only the 4 AGI sub-agents (agi-ceo,...
- fix: hot reload plugin management changes — Summary - add Gateway plugin hot reload for install/enable/disable/list/search management paths while keeping source-changing installs restart-backed - stop affected channel plugins before registry...
- fix(gateway): bound websocket auth after valid connect — Repair contributor PR #49751 in place. Plan: - Rebase the existing branch onto current main 381c2e1d1a3bbc5c41383e6d66b094bbdf3a6bf0. - Keep clearing the preauth timer only after a valid connect...
- [Bug]: Multiple regressions in 2026.4.8 - Gateway high CPU, CLI timeout, Discord voice broken, Messenger bot broken
- fix(tts): honor short explicit tagged speech text — Summary - honor explicitly tagged [[tts:text]]...[[/tts:text]] content even when it is short - keep the existing short-text guard for untagged auto-TTS replies - add regression coverage for both...
- [Bug]: TTS directives ( + [[tts:text]]) silently dropped on claude-cli backend; no voice delivere — Bug type Regression (worked before, now fails) Beta release blocker No Summary On openclaw 2026.4.26 with the claude-cli backend, model output containing and [[tts:text]]…[[/tts:text]] is parsed but...
- iMessage monitor: no per-turn watchdog for tool activity during agent silence — Symptom When an agent is silent past a threshold but is still running tools, the iMessage monitor sends nothing — the user has no signal that work is in progress. On longer turns this looks...
- Elevated exec still flaps after PR #74666 (v2026.4.29) — Summary Elevated exec permissions drop after the first approved command completes in the same assistant turn. PR #74666 ("preserve turnSourceChannel as messageProvider in approval followup runs") was...
- [Bug]: 2026.4.29 Web UI chat bubbles then disappears; direct OpenAI API works — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Hi OpenClaw Support, I’m having an issue where OpenClaw loads normally, but the Web UI chat does not...
- Task registry maintenance can burn CPU quadratically with many tasks — Summary OpenClaw 2026.4.29 can spend sustained CPU in task registry maintenance when the task registry contains many records. What I observed While validating an owned 2026.4.29 build on macOS, the...
- OpenClaw sponsors (123 total) — 123 sponsors supporting OpenClaw
- fix: add package identity repair admin
- fix: allow JSON Schema manifests in package publish
- fix: keep beta plugin packages off latest
- fix: add package owner transfer repair
- feat: reserve OpenClaw plugin package names
- Update securityPrompt.ts
- Update securityPrompt.ts
- Update securityPrompt.ts
- fix: rely on JSON artifact neutralization
- Update securityPrompt.ts
- fix: wrap ClawScan skill artifacts in prompt boundary
- ci: expand pr validation coverage — Split PR validation into explicit static, unit, package, type/build, HTTP e2e, and browser-smoke gates. Add local ci:* scripts and document the required status checks.
- fix: keep probe helper types private
News
- OpenClaw on DigitalOcean — 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...
- Google restricting Google AI Pro/Ultra subscribers for using OpenClaw
- OpenClaw is a security nightmare dressed up as a daydream
- OpenClaw is changing my life
- OpenClaw’s memory is unreliable, and you don’t know when it will break
- This is the most capable, highest level of sophistication agentic framework in the world right now. — I open sourced it yesterday. What makes that sentence even stranger is this. The agent itself built and published its own repository on…Continue reading on Medium »
- Testing Documentation Like Software in an Air-Gapped Environment — Most documentation does not fail because the writer was careless.Continue reading on Medium »
- MCP Is Not an API Layer. It’s an Agent Economy. — The industry keeps describing Model Context Protocol (MCP) as “agents calling tools.” That framing is convenient but dangerously…Continue reading on Medium »
- Running OpenClaw in a Docker Container on Ubuntu 22.04 — If you desire, you can follow the documentation that OpenClaw has proposed themself. You can find this documentation here.Continue reading on Medium »
- Hermes vs. OpenClaw: Which Open-Source AI Agent Really Wins? — 127K stars in 9 weeks. Here’s why Hermes is the fastest-growing AI agent on GitHub — and how it differs from the 345K-star veteran…Continue reading on Medium »
- OpenClaw, AI AGENT로 24/7 목표를 향해 스스로 움직이는 공장 — “ERP부터 MES까지를 AI가 관통하는 자율 생산 시스템”Continue reading on Medium »
- Claude Is Down Again? Here’s The Solution — Claude has been going down a lot lately. So much that it only has one 9 of uptime.Continue reading on Medium »
- My OpenClaw agent didn’t get dumber — I just gave it 50 skills and hoped for the best — A lot of “my OpenClaw agent got worse” stories start the same way: somebody stuffed 50 overlapping skills into one agent and called it…Continue reading on Medium »
- The Uniformed Guard Problem: Why AI Agent Sandboxes Need Identity, Not Just Policy — NemoClaw is NVIDIA’s reference stack for running OpenClaw agents safely. It wraps the agent in an OpenShell sandbox with a deny-by-default…Continue reading on Highflame »
- Your Hermes Agent Has No Performance Review. Here’s How to Fix That — Everyone is excited about the self-learning loop. Nobody is talking about what happens when it learns the wrong thing. Rollback strategies…Continue reading on Medium »
- How OpenClaw Agents Bypass Bot Detection (And How to Stop Them) (4 minute read) — May 01|InfosecHow OpenClaw Agents Bypass Bot Detection (And How to Stop Them) (4 minute read)
Don't miss what's next. Subscribe to Openclaw Newsletter: