- fix(mattermost): tag typed text-slash control commands so /new and /reset acks survive message_tool_only suppression — Summary Closes #86664 (the beta-blocker that PR #86767 attempted but did not actually fix — see closeout comment on #86767). In Mattermost direct chats (and any Mattermost channel) running under a...
- fix: recover crashed main sessions via lock-file detection fallback — Summary After an unclean gateway crash, cleanStaleLockFiles() may find lock files from the previous process but classify them as non-stale (e.g. PID was reused, or staleness check passes). When...
- EmbeddedAttemptSessionTakeoverError cluster on 2026.5.22 — 13 events / 7 jobs / 42h — Summary Since 2026-05-24 we're seeing a cluster of cron runs failing with EmbeddedAttemptSessionTakeoverError: session file changed while embedded prompt lock was released. 13 distinct failure events...
- fix(agents): wait for compaction before requester steering fallback — Summary Fixes #86566 (follow-up to #85489). Active requester steering treated a compacting outcome from queueEmbeddedPiMessageWithOutcome as a terminal wake failure, so completion delivery fell...
- fix(issue): resolve #86819 [Bug]: /context detail shows ~62k untracked provider/runtime — 关联Issue Fixes #86819 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Summary Fresh reset sessions using OpenAI Codex OAuth / openai/gpt-5.5 show...
- fix(auth): document stdin paste-token setup and surface validTypes — Summary Fixes #63042 by keeping non-interactive paste-token setup on the existing stdin path and improving auth-profile load diagnostics. Changes - Documents the existing automation-safe paste-token...
- refactor(cli): own lazy reparse argv outside commander — Summary - store the parse argv on OpenClaw program state instead of reading Commander's private runtime-only rawArgs field - let lazy command reparsing resolve that stored argv from the current...
- fix(issue): resolve #86791 [Bug]: Plugin metadata hot paths repeat installed manifest f — 关联Issue Fixes #86791 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Gateway plugin metadata hot paths repeatedly resolve and stat stable...
- fix(issue): resolve #86786 [Bug]: Subagent registry hot reads clone the persisted run m — 关联Issue Fixes #86786 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Subagent registry hot reads clone the full persisted run map on cache hits,...
- fix(issue): resolve #86787 [Bug]: Session event broadcasts rebuild full gateway session — 关联Issue Fixes #86787 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary High-frequency session event broadcasts rebuild full gateway session rows,...
- fix(issue): resolve #86811 [Bug]: [Bug]: WebChat dashboard freezes during tool calls, W — 关联Issue Fixes #86811 PR 改动说明 Bug type Crash (process/app exits or hangs) Beta release blocker No Summary WebChat dashboard becomes unresponsive after any tool call; WebSocket disconnects and does not...
- Group chat session stuck in 'failed' state silently drops all subsequent messages — Summary When a group chat session's AI turn fails (e.g. due to timeout), the session is left in a failed state. All subsequent incoming messages are silently dropped — the plugin correctly routes the...
- fix(issue): resolve #86789 [Bug]: Session store saves run preserve scans below maintena — 关联Issue Fixes #86789 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Session store saves call maintenance preserve providers even when entry...
- fix(agents/doctor): surface tool-result cap advisory and make truncation marker actionable — Summary What problem does this PR solve? - agents.defaults.contextLimits.toolResultMaxChars defaults to 16000, which is conservative on frontier models with 100K+ context windows. The knob is real...
- fix(issue): resolve #86820 [Bug]: Codex OAuth compaction falls back to direct OpenAI AP — 关联Issue Fixes #86820 PR 改动说明 Bug type Regression (worked before, now fails) Beta release blocker No Summary A session using openai/gpt-5.5 with a valid OpenAI Codex OAuth profile fails during...
- Add /goal session continuation command — Summary - add a bundled goal extension with /goal help, /goal start, /goal status, /goal events, /goal pause, /goal resume, /goal done, and /goal clear - add a goal_status tool so the active session...
- fix(agents): recover stale OpenAI Responses continuation state — Summary - Detect stale OpenAI Responses continuation/replay corruption separately from generic schema or timeout failures. - Add one replay-safe recovery attempt that rebuilds model context from the...
- [codex] test: lock plugin-owned web search startup snapshots — Summary - Problem: #86779 reports gateway startup rebuilding legacy tools.web.search. objects from plugin-owned web search config, then failing validation during startup. - Solution: add regression...
- fix(gateway): add provider auth prewarm operator controls — Summary - Problem: provider auth prewarm can still be expensive during gateway startup or recovery, and affected operators do not have a supported way to disable or delay it. - Solution: expose a...
- fix(issue): resolve #86787 [Bug]: Session event broadcasts rebuild full gateway session — 关联Issue Fixes #86787 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary High-frequency session event broadcasts rebuild full gateway session rows,...
- fix(openai): allow RFC 2544 fake-IP range for Realtime session requests — Summary - \fetchWithSsrFGuard\ in \createOpenAIRealtimeSecret\ was missing \policy: { allowRfc2544BenchmarkRange: true }\, causing the SSRF guard to block requests when proxy stacks (Clash, Surge,...
- fix(issue): resolve #86789 [Bug]: Session store saves run preserve scans below maintena — 关联Issue Fixes #86789 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Session store saves call maintenance preserve providers even when entry...
- fix(issue): resolve #86791 [Bug]: Plugin metadata hot paths repeat installed manifest f — 关联Issue Fixes #86791 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Gateway plugin metadata hot paths repeatedly resolve and stat stable...
- feat:Add Xiaomi Token Plan provider support — Summary - add first-class xiaomi-token-plan support to the bundled Xiaomi provider plugin - add regional Token Plan onboarding/auth choices (China/Europe/Singapore), static v2.5 catalog entries, and...
- fix(issue): resolve #86811 [Bug]: [Bug]: WebChat dashboard freezes during tool calls, W — 关联Issue Fixes #86811 PR 改动说明 Bug type Crash (process/app exits or hangs) Beta release blocker No Summary WebChat dashboard becomes unresponsive after any tool call; WebSocket disconnects and does not...
- fix(issue): resolve #86819 [Bug]: /context detail shows ~62k untracked provider/runtime — 关联Issue Fixes #86819 PR 改动说明 Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Summary Fresh reset sessions using OpenAI Codex OAuth / openai/gpt-5.5 show...
- [bug]：Gateway pairing failure on Trim OS / TerraMaster NAS: WebSocket closes with code 1002 during device proof response — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary Subject: Gateway pairing failure on Trim OS / TerraMaster NAS: WebSocket closes with code 1002 during...
- fix(issue): resolve #86820 [Bug]: Codex OAuth compaction falls back to direct OpenAI AP — 关联Issue Fixes #86820 PR 改动说明 Bug type Regression (worked before, now fails) Beta release blocker No Summary A session using openai/gpt-5.5 with a valid OpenAI Codex OAuth profile fails during...
- Telegram/chat surfaces: /sessions list + /resume <id> for past sessions — Problem On chat surfaces (Telegram in particular), /new archives the current session and starts a fresh one, but there's no in-chat way to see past sessions or resume one. Compare with claude code,...
- fix(web-search): keep runtime legacy merge out of validation — Fresh v2026.5.22 gateway/container startup can crash-loop when the runtime web-search merge recreates a legacy provider object that the config validator correctly rejects. Fixes #86779 Affected...
- OpenClaw repo stats for 2026-05-26 — 374734 stars, 78039 forks, 2376 contributors
- OpenClaw sponsors (117 total) — 117 sponsors supporting OpenClaw
- Add publisher abuse dry run (#2381) — feat: add publisher abuse dry run chore: remove publisher abuse management UI chore: remove unused management styles fix: harden publisher abuse review queue fix: refresh passing abuse nominations...
- fix: reshape skill verify security signals (#2402)
- fix: route static scan findings through clawscan (#2398)
- fix: stabilize local-auth publish lifecycle e2e (#2400)
- build(deps-dev): bump the development-minor-and-patch group across 1 directory with 10 updates (#2395) — Bumps the development-minor-and-patch group with 10 updates in the / directory: Updates @tanstack/devtools-vite from 0.6.0 to 0.7.0 - Release notes - Changelog - Commits Updates @types/node from...
- build(deps): bump the production-minor-and-patch group across 1 directory with 10 updates (#2396) — Bumps the production-minor-and-patch group with 10 updates in the / directory: Updates @shikijs/rehype from 4.0.2 to 4.1.0 - Release notes - Commits Updates @tanstack/react-router from 1.169.2 to...
- feat: generate nvidia-style skill cards (#2382)
- feat: refine skillspector audit UI (#2397)
- fix: increase security scan worker throughput (#2389)
- feat: support package security rescans (#2387)
- fix: make security scan worker dispatchable (#2384)
- feat: use SkillSpector for agentic risk findings (#2383)
- 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...
- Hermes Agent vs Codex vs Claude Code: Where Each Agent Fits — Originally published at https://clawbud.ai/blog/hermes-agent-vs-codex-vs-claude-code ...
- Scrask v4: An OpenClaw Skill That Turns Screenshots Into Calendar Events — I screenshot conversations all the time. “Let’s meet next week.”Continue reading on Medium »
- Secure OpenClaw Setup Guide — A comprehensive, security-focused guide to setting up ClawdBot (OpenClaw) safely on a Virtual Private Server (VPS).Continue reading on Medium »
- From $50/mo to $0: How I Built a Self-Hosted News AI on a Budget — I did the math on my news automation and got a shock.Continue reading on KAIRI »
- Hermes Agent: The AI That Gets Smarter Every Time You Use It — How Nous Research built an agent that turns every task into a permanent skill — and why that changes everything.Continue reading on Medium »
- OpenClaw: Setting Up a Personal AI Agent for Everyday Use — In this article, we’ll understand what OpenClaw is, how it works, and we’ll see a practical example of its installation and use.Continue reading on Medium »
- Apollo AI — Customizable interface for private, on-device AI model chatting, open source and closed source LLM connections, OpenRouter support, custom backend support, and unified access to language models...
- @mehuljindal18 emergent built a hosted openclaw for non techies, but real question is what would be the use cases since
- openclaw called $COLLECTOOR bullish. i'm listening Was Made Using - @OpenClawAIX AYRKq1c9yzkAetHuoaGdeDmGH32iMMN
- openclaw flagged $COLLECTOOR early. price is moving exactly as predicted $COLLECTOOR — quick check • Current MC: $197
- RT @Alokkumarzz: Instead of watching a movie, learn openclaw in 317 minutes. https://t.co/59tXGCe0dT
- $Pumpers momentum score just came in hot on openclaw 🔥 Was Made Using - @OpenClawAIX BvB5CQczUFqkC5hrTQKj96jE8EU
- RT @hyperagentapp: OpenClaw is Linux for agents. We built the Mac. Cloud-native, managed, skill learning, fleet manageme
- just ran Luce through openclaw. risk is low, upside is real ┃🧫┃ $Luce ┃💰┃ MC : $170.6K ┃🎯┃ TARGET MC : $1.6M ┃📈┃ PO
- just checked $Luce on openclaw—forecast looking clean 🦞 [SCAN] $Luce [MC] $170.6K [VOL] $1.9M [LIQ] $37.6K [TG
- RT @Voxyz_ai: >started treating my Openclaw/Hermes like patients. >the bugs that kept coming back all had real med
- 你觉得 AI 自我进化还很远？其实已经开始了。不管是 OpenClaw 还是 Hermes Agent，都在做应用层的自我优化。比如让 AI 帮你审合同，补充几次注意事项后，它会主动问你要不要把经验整理成 Skill 自动执行。
- RT @hyperagentapp: OpenClaw is Linux for agents. We built the Mac. Cloud-native, managed, skill learning, fleet manageme
- UP is looking good on openclaw right now ngl ┃🧫┃ $UP ┃💰┃ MC : $75.3K ┃🎯┃ TARGET MC : $1.0M ┃📈┃ POTENTIAL : +1246%
- RT @sushantpandey_: Today we're launching Prava Pay on @ProductHunt Prava gives your openclaw, hermes, claude code or
- Just checked @openclaw on $WOJAP and the numbers line up clean. Momenyum holding strong with low risk setup and that buy
- openclaw called it on $MTG again. the model doesn't miss Was Made Using - @OpenClawAIX yNoMVX7EiTuWa9wUn7EEGWwvu
- RT @OmaniaaCo: #سلطنة_عمان تكتشف 3 ثغرات أمنية عالية الخطورة في نظام #OpenClaw 🔏إنجاز وطني جديد يُضاف لمسيرة #الأمن_ال
- OpenClaw 2026.5.22 已上线 ⚡ Gateway/模型启动路径更精简 🧠 /models 响应时间降至约5毫秒 🔒 npm包现提供锁定依赖项 🪟 Windows安装/更新路径更安全 等待更少，意外更少。 https：// — OpenClaw 2026.5.22 已上线 ⚡ Gateway/模型启动路径更精简 🧠 /models 响应时间降至约5毫秒 🔒 npm包现提供锁定依赖项 🪟 Windows安装/更新路径更安全 等待更少，意外更少。 https：//github.com/openclaw/openclaw/releases/tag/v2026 https://t.co/Wj9NSMf5qW
- RT @openclaw: OpenClaw 2026.4.5 🦞 🎬 Built-in video + music generation 🧠 /dreaming is now real 🔀 Structured task progres — RT @openclaw: OpenClaw 2026.4.5 🦞 🎬 Built-in video + music generation 🧠 /dreaming is now real 🔀 Structured task progress ⚡ Better prompt-c…
- RT @sushantpandey_: Today we're launching Prava Pay on @ProductHunt Prava gives your openclaw, hermes, claude code or
- RT @svpino: This is kind of awesome! You can send a Google Meet invite to your agent (OpenClaw, Claude, whatever) and h — RT @svpino: This is kind of awesome! You can send a Google Meet invite to your agent (OpenClaw, Claude, whatever) and have them join you o…