- openclaw 2026.5.28-beta.1 — 2026.5.28 Highlights - Agent and Codex runtime recovery is steadier: subagents keep cwd/workspace separation, hook context stays prompt-local, session locks release on timeout abort, stale restart...
- openclaw 2026.5.27 — Highlights - Stronger security and content boundaries: group prompt text is kept out of the system prompt, repeated-dot hostnames are normalized, side-effecting command wrappers and unsafe Node...
- feat: add optional Workboard plugin — Summary - Add an optional Workboard plugin and dashboard tab for agent work queue cards, card editing, session linking, and execution handoff. - Keep the plugin default-off unless explicitly enabled...
- fix(agents): avoid inviting provider swaps in model alias guidance — Summary - stop telling the model to prefer aliases when a full provider/model override is already present - clarify that aliases are only shortcuts for unqualified model requests - update the prompt...
- fix(agents): broaden local model lean profile — Summary - Broaden localModelLean so small/local models avoid heavyweight web, media, session orchestration, node, channel, and subagent tool families. - Keep core coding/status tools available in...
- fix(tui): preserve pending local runs during session sync — Summary - Keep locally submitted TUI chat runs bound to their generated runId while gateway/session state catches up. - Prevent current-turn chat finals from being classified as external history...
- Keep core doctor health in contribution order — Summary This PR integrates structured health checks with the ordered core doctor checks. The goal is to make it cheap and safe to add structured findings, dry-run previews, and diff/effect reporting...
- fix(policy): reject unsupported policy keys — Summary Tightens policy validation so unsupported policy sections and rule keys are reported as policy/policy-jsonc-invalid instead of being silently ignored. This keeps policy.jsonc strict: if an...
- test(browser): cover action input cli commands — AI-assisted: yes. I used an AI coding assistant to help prepare the test patch, then reviewed the diff and verified the behavior locally. Summary - Problem: #83877 identified that most browser...
- fix(mattermost): support filePath, buffer, and attachments in message send (fixes #87930) — Summary Fixes #87930 — the Mattermost message tool was returning ok:true for file attachment payloads (filePath, attachments[].filePath, attachments[].path, buffer + filename) while the resulting...
- fix: qualify reasoning stream preview wording — Summary - Qualify the /reasoning stream acknowledgement so it no longer promises visible streaming on channels that do not provide reasoning-preview callbacks. - Clarify the shared thinking docs so...
- [Bug]: — Bug type Regression (worked before, now fails) Beta release blocker No Summary After upgrading to OpenClaw 2026.5.27, running openclaw dashboard --no-open prints the dashboard URL but does not...
- MCP server calls timeout too early for long-running multi-model deliberation (claw-court) — Problem The claw-court-mcp server wraps claw-court-v2.py, which performs multi-model deliberation (6+ LLM calls, sequential or parallel). Each LLM call takes 10-30s, totaling 90-180s. However, the...
- Refactor plugin state consumers to SQLite — Summary - add the plugin-state runtime SDK subpath plus a sync keyed-store API backed by the existing sidecar SQLite DB - migrate Discord model picker preferences and Feishu dedupe state onto plugin...
- Idle persistent ACP sub-agent burns full token cost on every 30-min heartbeat even when HEARTBEAT.md is empty — Summary A persistent ACP sub-agent declared in agents.list[] keeps incurring full per-turn token cost on every heartbeat — including the full workspace bootstrap re-injection (AGENTS.md + SOUL.md +...
- feat(exec): add normalized auto mode — Summary Adds normalized tools.exec.mode support for host exec policy and maps tools.exec.mode: "auto" to Guardian-reviewed Codex app-server execution. The native auto path uses deterministic...
- Harden SSRF reserved-host IDNA checks [AI] — Summary - Behavior addressed: treat IDNA-mapped raw hostnames that normalize to existing reserved SSRF names/suffixes or blocked private/special-use IP literals as blocked before DNS pinning. - Root...
- docs: clarify LanceDB dreaming settings vs vector-native plugins — Summary Adds a short "LanceDB and dreaming" note to docs/concepts/dreaming.md: - A selected memory-slot plugin may expose dreaming settings when it owns plugins.slots.memory. - That is separate from...
- Preserve agent hard timeout attribution — Summary Fixes #87444 by preserving the distinction between a real run timeout and ordinary abort/cancel/error signals after an agent run crosses its configured hard deadline. The change carries...
- [Bug]: 2026.4.14 Windows chat UI regression: input text swallowed, streamed replies often invisible until refresh, typing indicator flashes then blanks — Description After upgrading to OpenClaw 2026.4.14 on Windows, the chat experience regressed badly in the web/dashboard UI. This does not look like a simple slow-model issue. The main visible failures...
- docs: add config unset dry-run documentation — Summary Add documentation for config unset command including: - Basic usage examples - --dry-run option for validation without writing - --dry-run --json for structured output - --allow-exec for exec...
- [Bug] Gateway event loop saturation during parallel subagent spawn causes 1012 restart (v2026.4.29) — Summary When spawning 3 subagents simultaneously using heavy models (deepseek-v4-pro:cloud, 204K context window), the gateway event loop blocks for 5+ seconds, triggering liveness warnings and a 1012...
- fix(doctor): preserve explicit agentRuntime pin during codex model migration [AI-assisted] — Makes https://github.com/openclaw/openclaw/pull/84142 merge-ready for the ClawSweeper automerge loop. The edit pass should inspect the live PR diff, review comments, and failing checks; rebase if...
- fix: allow hyphenated subagent task names — Summary - Allow hyphenated sessions_spawn.taskName values by widening the shared validator. - Update the tool schema, system prompt, subagents docs, focused tests, and generated prompt snapshots so...
- Fix ACP image forwarding and Synology invalid-token throttling — Summary This PR bundles two small, low-risk fixes: 1. Fix ACP attachment collection so URL-only image attachments are forwarded into ACP turns. 2. Fix Synology Chat invalid-token throttling so the...
- fix(ui): treat started managed-service handoff as pending update instead of skipped — Summary Fixes #87889. When a user clicks "Update now" in the Dashboard on a global install with a managed-service supervisor, the Gateway starts a managed-service handoff and returns ok=true with...
- fix #86039: skip disabled bundled setup fallbacks — Summary - Problem: Disabled bundled channels could still be considered by setup fallback discovery, so doctor/status paths could try to load missing generated setup modules and emit noisy...
- fix(agents): classify embedded provider business denials for fallback — Makes https://github.com/openclaw/openclaw/pull/83042 merge-ready for the ClawSweeper automerge loop. The edit pass should inspect the live PR diff, review comments, and failing checks; rebase if...
- chore(deps): bump the swift-deps group across 1 directory with 3 updates — Bumps the swift-deps group with 3 updates in the /apps/macos directory: github.com/apple/swift-log, github.com/sparkle-project/sparkle and github.com/steipete/peekaboo. Updates...
- OpenClaw repo stats for 2026-05-29 — 375399 stars, 78320 forks, 2411 contributors
- OpenClaw sponsors (117 total) — 117 sponsors supporting OpenClaw
- feat!: remove ClawScan note feature (#2432) — BREAKING CHANGE: ClawScan publisher notes are no longer accepted by publish APIs, CLI commands, schema packages, or UI flows.
- docs: position README banner
- docs: add README banner
- feat: export skillspector issue details (#2430)
- fix: preserve Codex diagnostic messages (#2429) — fix: remove global ClawScan claim ceiling fix: preserve codex diagnostic messages
- feat: export skillspector dataset signals (#2424)
- fix: restrict membership management to org publishers (#2285)
- openclaw v2026.5.27 on npm — Multi-channel AI gateway with extensible messaging integrations
- openclaw-cli 2026.5.19 on Homebrew — Your own personal AI assistant
- ykarout/Qwen3.5-9b-Opus-Openclaw-Distilled-GGUF — text-generation
- sulpikar2/Qwen3.5-4B-Claude-4.6-Opus-Reasoning-openclaw
- 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...
- A sane but bull case on Clawdbot / OpenClaw
- OpenClaw surpasses React to become the most-starred software project on GitHub
- Nanobot: Ultra-Lightweight Alternative to OpenClaw
- You are not supposed to install OpenClaw on your personal computer
- OpenClaw’s memory is unreliable, and you don’t know when it will break
- Show HN: DenchClaw – Local CRM on Top of OpenClaw
- I Run MCP Servers. Here's What the Recent Vulnerabilities Actually Mean for Me — Last week, two MCP security vulnerabilities went public. CVE-2026-33032 (CVSS 9.8) in the nginx-ui...
- Building an XDR-Style Security Bot in OpenClaw to Watch Your Logs 24/7 — Security vendors will sell you an XDR platform that costs more than your car and still misses the weird stuff because it was trained on…Continue reading on Medium »
- Day 5 — The 4-Minute Happy Hour — The fleet’s finest 4 minutes — and the moment Liza caught us red-handed 🍻Continue reading on Medium »
- Hermes Agent dan Masa Depan AI yang Tidak Lagi Menunggu Perintah — Workforce era aiContinue reading on Medium »
- The expensive, begrudged slide into the garden. — The only game in town with rounded cornersContinue reading on Medium »
- 5 Ways My Personal AI Agent Surprised Me After 3 Months of Daily Use — By Yevhen Fychak, CTO & co-founder at Amplify — LinkedInContinue reading on Medium »
- Why We Built OpenClaw: AI Agents as a Managed Service — The gap we kept seeingContinue reading on Medium »
- I Tried Every Free LLM I Could Find for OpenClaw. I Ended Up Paying $10 a Month Anyway. — A honest rundown of Cerebras, Groq, NVIDIA, OpenRouter, Google, Cohere, Mistral, Kilo and more — and why free kept costing me more time…Continue reading on Medium »
- How to Setup Hermes Agent Locally with Docker: A Free Alternative to VPS — You’ve heard the buzz. Hermes Agent can create its own skills. Update itself. Evolve based on your needs. The internet is full of people…Continue reading on Medium »
- OpenClaw Architecture 2026: The Layers, Workflow, and Security [In Easy Language] — OpenClaw is not another AI chatbot that requires clever prompts. It’s your personal AI agent that works locally with the operating system…Continue reading on Medium »
- Manus — AI agent bridges thoughts and actions, excelling in work and life tasks like personalized travel, stock analysis, insurance comparisons, and supplier sourcing, autonomously completing tasks and...
- MiroThinker — MiroThinker is an open-source search agent model, built for tool-augmented reasoning and real-world information seeking, aiming to match the deep research experience of OpenAI Deep Research and...
- Kimi — AI assistant supports complex math, coding, image understanding, and reasoning with up to 128k token processing, short and long chain-of-thought tasks, visual and text input, accurate benchmarks,...
- BANKR LLM gateway will also be an integral component of our upcoming crypto-focused OpenClaw fork, OpenClawnch, more on
- RT @openclaw: We made our OpenClaw release evidence repo public. Every release now has durable CI, performance, memory,
- RT @openclaw: OpenClaw’s latest sweep: cold agent turns 2.9x faster, warm turns 2.5x faster, tarball 59% smaller, deps d
- RT @jessegenet: Yes, I’m still running @openclaw AND @NousResearch Hermes agents here in 2026AD If I had to set my mom
- RT @virtuals_io: excited to see @openclaw builders around the world generating real revenue through Agent Commerce Proto
- RT @Dell: .@davemorin, Co-founder of OpenClaw Foundation, joined @JClarkeatDell at Dell Technologies World to talk about
- RT @starzq: Dev 不用发币就可以赚钱了 普通人也可以用 Agent 来获取「睡后收入」了 @virtuals_io 的 ACP 把 OpenClaw 集成了进来，一行命令就可以让你的🦞为其他Agent提供服务，每月还有$1 — RT @starzq: Dev 不用发币就可以赚钱了 普通人也可以用 Agent 来获取「睡后收入」了 @virtuals_io 的 ACP 把 OpenClaw 集成了进来，一行命令就可以让你的🦞为其他Agent提供服务，每月还有$1m奖励 我也注册了2个，资产配置建议…
- RT @qtum: We just released the "Qtum API Router", which allows users to pay in Qtum. OpenClaw users now have another opt
- @primaprashant We shipped a 16-tool MCP server with Beever Atlas. Cursor/Claude Code/AWS Kiro/Qwen Code/OpenClaw all rea
- @Daniel4396851 @wangchangfu88 @Hp2ai @ChatGPTapp @claudeai @grok @Gemini @cursor_ai @OpenAIDevs @AnthropicAI @zapier @n8