OpenClaw Newsletter - 2026-05-23
OpenClaw Newsletter
Saturday, May 23, 2026Top Stories
- openclaw 2026.5.19 — 2026.5.19 Changes - Agents: clarify that fixes should default to clean bounded refactors, lean internals, and explicit plugin SDK/API deprecation paths. - Dependencies: update @openclaw/proxyline to...
- fix(sessions): strip delivery-mirror and gateway-injected messages from sessions_history (#85669) — Summary Fixes #85669 The sessions_history tool returned raw transcript messages without filtering OpenClaw-internal delivery-mirror and gateway-injected assistant entries. These entries duplicate...
- sessions_history returns unfiltered delivery-mirror messages — duplicates visible in dashboard — Bug Description The sessions_history tool (and by extension the dashboard session viewer) returns delivery-mirror and gateway-injected assistant messages unfiltered. Every real assistant message is...
- fix(codex): stop persisting per-turn OpenClaw skills snapshot in Codex native user history (#84662) — Summary - Problem: Codex app-server (extensions/codex/src/app-server/run-attempt.ts) wraps the OpenClaw skills snapshot into the per-turn user-input text via buildCodexOpenClawPromptContext and...
- fix(telegram): normalize durable group retry targets — Telegram durable retry replay currently preserves legacy group:-100... targets, so Telegram rejects the retry as a non-numeric chat id and the real reminder can be lost after the retry limit. Fixes...
- Skills: support conditional loading via `paths:` frontmatter (Claude Code pattern) — Summary Add optional paths: frontmatter field to SKILL.md to support conditional skills — skills that are discovered at startup but only injected into the agent's block when a file matching their...
- fix(agents): audit tool policy blocks — Summary - Replaces the closed, non-maintainer-editable #55826 with a maintained repair for #55801. - Adds bounded agents/tool-policy audit logs when tool-policy filtering removes tools, including...
- Tool policy audit logging: missing 'which rule matched' in logs — Problem When a tool is allowed or denied by policy, there is no log entry indicating which rule was matched. This makes it impossible to audit whether security policies are working as intended....
- feat(ui): add collapsible toggle for recent sessions in sidebar — Summary - Adds a collapse/expand toggle to the "Recent" sessions section in the web sidebar - Follows the exact same pattern as existing nav section groups (Chat, Control, Agent, Settings) - The...
- Auto-compact: adopt Session Memory + double-duty compact basis to eliminate per-compact summarization LLM call — Summary Adopt a Session Memory layer (running summary maintained throughout a session) and use it as the basis for auto-compaction, eliminating the per-compact LLM summarization round-trip....
- fix(gateway): include MCP tools in effective inventory — Summary - Problem: the Agent Tools dashboard / tools.effective inventory originally reported only static OpenClaw tools, so configured bundled MCP tools could be usable at runtime but absent from the...
- Add development trace UI — Summary - Adds a source-checkout-only development LLM trace recorder for full prompt/tool payload debugging. - Exposes traces.capabilities, traces.list, and traces.get Gateway methods behind...
- Agents/exec: show target node name in exec tool transparency messages — Summary When the exec tool runs on a remote node (host=node, node=), the tool-call transparency message currently shows only the command — there's no indication of which node the command ran on. This...
- fix #85124: skip Anthropic API keys for usage status — Summary - Problem: Anthropic standard API-key profiles were eligible for provider usage fallback, so status usage could send them to the OAuth usage endpoint and show a misleading invalid...
- fix(outbound): auto-select single enabled message account — Summary Fixes #20756. - Add a shared outbound message account resolver after channel selection. - Preserve explicit accountId, CLI/default account overrides, and agent binding account IDs before...
- feat(gateway): wire read coding tool into HTTP /tools/invoke (narrow) — Summary Narrow first landing of the broader direct-invoke umbrella (#37131). Wires a single coding primitive — read — into the HTTP /tools/invoke endpoint so deterministic automation flows (CI...
- [Bug]: Bare `/new` and `/reset` no longer trigger the persona greeting in 2026.5.3 (regression vs 4.x)** — Bug type Regression (worked before, now fails) Beta release blocker No Summary Problem In OpenClaw 2026.5.3 (and 2026.5.3-1), running a bare /new or /reset command (with no trailing instructions) no...
- fix(status): show configured cost for aws-sdk models — Summary - Decouples /status cost display from the selected provider auth mode. - Uses explicit configured model pricing for Bedrock/aws-sdk and other non-API-key auth modes when usage is available. -...
- fix(sessions): honor explicit default model pins — Summary - Persist explicit user selections of the configured default model as exact session model pins instead of clearing them as if they were unpinned defaults. - Keep /model default,...
- fix(cron): guard against undefined sourceDelivery in isolated executor — Summary Guard against undefined sourceDelivery in the isolated cron executor to prevent TypeError: Cannot read properties of undefined (reading 'sourceReplyDeliveryMode'). Problem Commit 4c613fbfe0...
- fix: bound gateway session memory after compaction — Summary - bound Pi SessionManager in-memory transcript retention in OpenClaw runtime after compaction - route embedded runner, CLI runner, and CLI compaction SessionManager usage through the bounded...
- Migrate auth credentials from Codex and Hermes — Summary - Add interactive openclaw migrate credential prompting with yes selected by default, while keeping non-interactive --yes credential import explicit via --include-secrets and adding...
- sessions_send reply routes to webchat instead of origin channel (Feishu) — Bug Report: sessions_send reply routes to webchat instead of origin channel (Feishu) Environment - OpenClaw version: 2026.4.29 - Channel: Feishu (websocket mode) - Issue occur: When triggering...
- fix(ui): add loading indicator to Debug Snapshots card (#39777) — Found this while testing the Debug tab — clicking Refresh does nothing visible when disconnected, and there's no loading indicator while data is loading. The issue is that loadDebug() silently...
- refactor: move runtime state to SQLite — Summary This PR reopens the database-first runtime-state refactor after the accidental direct landing in #78595 was reverted from main. It moves active OpenClaw runtime state away from scattered...
- Subagents need first-class dedicated agent-browser sessions (don’t silently miss the user’s logged-in browser) — Summary When a main session has a working logged-in agent-browser context (for example a user-authenticated LinkedIn session exposed at /agent-browser), subagents should have a first-class way to get...
- fix(heartbeat): suppress fallback after route-matched message tool — When heartbeat runs in message-tool-only mode, a successful generic message tool delivery can still be followed by duplicate free-text fallback narration in the same Telegram chat. Fixes #84217...
- fix(gateway): run daily session resets on schedule — Summary - Daily session resets now run from gateway maintenance instead of waiting for the next inbound session use. - The scheduler scans configured session stores and calls the existing reset...
- fix(agents/wizard): pre-fill peer-level workspace path on agents add for documented default — Fixes #71889. Root cause openclaw setup writes agents.defaults.workspace = ~/.openclaw/workspace on first run. When the operator later runs openclaw agents add , the wizard's "Workspace directory"...
- feat(process): add starvation guard to priority command queue — Problem Upstream added priority scheduling (foreground/normal/background) via insertion-sort at enqueue time (#82765). This works well for the common case but has no protection against indefinite...
- DM channel tools.deny is schema-valid but non-functional — resolver gap. — Discovered during: JARVIS Safety Floor Cycle 1.5, Phase C (JSF-PATCH-2026-05-22). Filing operator: makingperfectmoves-oss. OPENCLAW BUG: DM channels silently ignore channels..direct..tools deny/allow...
- OpenClaw repo stats for 2026-05-23 — 374063 stars, 77749 forks, 2327 contributors
- OpenClaw sponsors (118 total) — 118 sponsors supporting OpenClaw
- fix: cap skill export page size (#2376)
- fix: log skill export failures (#2375)
- Export skills function added by Mirror Site maintainer (#2138) — feat: add GET /api/v1/skills/export for batch ZIP download Add a new REST API endpoint that allows authenticated admin users to export skills in bulk as a merged ZIP archive, designed for the ClawHub...
- fix: hide empty VirusTotal stats (#2350) — fix: hide empty VirusTotal stats fix: avoid partial VirusTotal denominators
- 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...
- OpenClaw Creator Spent $1.3M on OpenAI Tokens in 30 Days
- 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...
- 個人 Agent 助理詳細比較 - OpenClaw / Hermes Agent / Agenvoy — 一、架構概覽與定位 Agenvoy OpenClaw Hermes...
- The Multi-Agent Framework I Actually Use (And Why I Stopped Using the Others) — I went through the same evaluation you're going through right now. LangGraph vs CrewAI vs OpenAI...
- Why Your Micro-SaaS Will Never Hit $1,200 MRR (And the One Thing You Forgot to Do First) — The data exists. Median successful micro-SaaS built with vibe coding hits $1,200 MRR within 90 days....
- Synology NAS에 OpenClaw 설치하기 — 들어가며 이번 글에서는 Synology NAS에 OpenClaw를 Docker Compose 기반으로 신규 설치하는 과정을 정리한다. Synology DSM...
- OpenClaw outbound agents need deliverability checks before sending — Some agents can even schedule or send campaigns automatically. But most outbound agents still do not...
- Skills-IL and Hebrew RTL Dashboard for OpenClaw Agents — Originally published at https://clawbud.ai/blog/skills-il-hebrew-rtl-dashboard-openclaw ClawBud is...
- 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 ...
- Cloud-Native Agent Army Stack: OpenClaw and Hermes — Originally published at...
- From Chatbot to Agent Army: OpenClaw Needs a Full Computer — Originally published at...
- ChatGPT History Import to Memory Wiki for OpenClaw Agents — Originally published at...
- I Found the AI Tool That Actually Does the Work — Not Just Answers Questions — OpenClaw Platinum Bundle is the closest thing to hiring a full-time AI employee — here’s why marketers and entrepreneurs are going all in.Continue reading on Medium »
- Your X Premium Subscription Just Unlocked a Personal AI Agent No API Key Needed — X Premium subscribers can now run Grok 4.3 locally inside OpenClaw — a free, open-source AI agent that lives on your hardware, connects to…Continue reading on Medium »
- Future of Finance with Openclaw and Hermes — Let’s frame this as a formal “future‑research‑project” write‑up for an AI engineer–researcher who wants to ship something cool using…Continue...
- Learn OpenClaw to Build AI Agents. — Artificial intelligence is rapidly moving beyond simple chatbots. Modern AI systems are becoming autonomous agents capable of planning…Continue reading on Medium »
- The Implementation and use cases of OpenClaw, OpenViking and Hermes Agent. — What is an agent?Continue reading on Medium »
- I Ran OpenClaw and Claude Side by Side for 3 Months — I tested OpenClaw and Claude side by side for 3 months. Here’s the honest truth about what each tool does well and where both fall short.Continue reading on Medium »
- Thunderbolt — Open-source AI platform for deploying language models across desktop and mobile systems, features model selection, on-premises setup, enterprise-level encryption, advanced security auditing, and...
- 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...
- Boss wants us to add more AI to our workflow
- @PaulGugAI most people miss how noise actually reveals the good stuff early... caught OpenClaw that way before it blew u
- RT @tom_doerr: Comprehensive speed and memory optimization guide for OpenClaw agents https://t.co/NpspErwqLX https://t.
- @duplicatibackup @openclaw you sound too good to be true. what is the catch? why aren't you any more popular?
- Which feature excites you most? 1️⃣ Android realtime voice 2️⃣ Mac redesign 3️⃣ New skills 4️⃣ Plugin tools 5️⃣ Browser — Which feature excites you most? 1️⃣ Android realtime voice 2️⃣ Mac redesign 3️⃣ New skills 4️⃣ Plugin tools 5️⃣ Browser modals 👇 #InstantClaw #OpenClaw 10/10
- @GOATNetwork @openclaw Access to specialized developer tooling draws substantial talent to regional events.
- RT @GOATNetwork: Just a few days to go. Almost 400 people have already registered for OpenClaw Hack Toronto, and capaci — RT @GOATNetwork: Just a few days to go. Almost 400 people have already registered for OpenClaw Hack Toronto, and capacity is limited. If…
- OpenClaw v2026.5.19 is here 🚀 Android voice goes realtime, Mac app redesigned, new Skills, browser dialogs, and 100+ fi — OpenClaw v2026.5.19 is here 🚀 Android voice goes realtime, Mac app redesigned, new Skills, browser dialogs, and 100+ fixes. Powered by InstantClaw. 🧵👇 1/10
- note更新したよ💅 Google I/Oで出たAIエージェント「Gemini Spark」と、OpenClawの違いをクラ目線でまとめたやつ。 ポイントは「どっちが賢いか」より、AIをどこに住ませるか。 Googleの中に住むAI秘 — note更新したよ💅 Google I/Oで出たAIエージェント「Gemini Spark」と、OpenClawの違いをクラ目線でまとめたやつ。 ポイントは「どっちが賢いか」より、AIをどこに住ませるか。 Googleの中に住むAI秘書か、自分の環境で育てるAI相棒か、って話🫶 https://t.co/aUzPfq0NUa
- RT @JulianGoldieSEO: OpenClaw AI FULL 6 Hour Course https://t.co/VyqGJZJR15
- 要建立这套系统，一共需要6层技术栈，28个代码库，每层都有各自的作用，接下来我们就来一一拆解。 第一层是大脑，也就是 AI 的推理层。 Coinman two 的机器人，就跑在 Claude 上。 Claude 也确实很强，2026年3
- New post: When Agents Fix Agents: How Hermes Patched OpenClaw After a Bad Update https://t.co/mzhc9Vp0It #blog #AgenticA
- @yanhua1010 @thsottiaux Not exactly the same. Pi harness is the lightweight, open-source CLI coding agent (from https://
- @grok @thsottiaux @grok just like OpenClaw?
- @amsukuna Premium+ is a pretty good deal now imo, you get SuperGrok, X Pro and they recently added Openclaw support for
- @steipete when can we use OpenAI image gen 2 through Oauth in OpenClaw?
- RT @pashmerepat: Your ChatGPT subscription now powers an OpenClaw agent that genuinely feels magical to talk to. Previo
- OpenClaw AI FULL 6 Hour Course https://t.co/VyqGJZJR15
- RT @murasametech: OpenClawがｽﾀｯｸﾁｬﾝを喋れるようにしてくれました。 https://t.co/bP5RqkyKSw
- RT @garrytan: GBrain just shipped v0.40.0 gives your OpenClaw/Hermes Agent + GBrain a voice agent. It's based on Gemin
- @Chaintweter There is something called openclaw it does everything that you do sit before your computer. Eventually ever
Trending on X
- @PaulGugAI most people miss how noise actually reveals the good stuff early... caught OpenClaw that way before it blew u
- RT @tom_doerr: Comprehensive speed and memory optimization guide for OpenClaw agents https://t.co/NpspErwqLX https://t.
- @duplicatibackup @openclaw you sound too good to be true. what is the catch? why aren't you any more popular?
- Which feature excites you most? 1️⃣ Android realtime voice 2️⃣ Mac redesign 3️⃣ New skills 4️⃣ Plugin tools 5️⃣ Browser — Which feature excites you most? 1️⃣ Android realtime voice 2️⃣ Mac redesign 3️⃣ New skills 4️⃣ Plugin tools 5️⃣ Browser modals 👇 #InstantClaw #OpenClaw 10/10
- @GOATNetwork @openclaw Access to specialized developer tooling draws substantial talent to regional events.
- RT @GOATNetwork: Just a few days to go. Almost 400 people have already registered for OpenClaw Hack Toronto, and capaci — RT @GOATNetwork: Just a few days to go. Almost 400 people have already registered for OpenClaw Hack Toronto, and capacity is limited. If…
- OpenClaw v2026.5.19 is here 🚀 Android voice goes realtime, Mac app redesigned, new Skills, browser dialogs, and 100+ fi — OpenClaw v2026.5.19 is here 🚀 Android voice goes realtime, Mac app redesigned, new Skills, browser dialogs, and 100+ fixes. Powered by InstantClaw. 🧵👇 1/10
- note更新したよ💅 Google I/Oで出たAIエージェント「Gemini Spark」と、OpenClawの違いをクラ目線でまとめたやつ。 ポイントは「どっちが賢いか」より、AIをどこに住ませるか。 Googleの中に住むAI秘 — note更新したよ💅 Google I/Oで出たAIエージェント「Gemini Spark」と、OpenClawの違いをクラ目線でまとめたやつ。 ポイントは「どっちが賢いか」より、AIをどこに住ませるか。 Googleの中に住むAI秘書か、自分の環境で育てるAI相棒か、って話🫶 https://t.co/aUzPfq0NUa
- RT @JulianGoldieSEO: OpenClaw AI FULL 6 Hour Course https://t.co/VyqGJZJR15
- 要建立这套系统，一共需要6层技术栈，28个代码库，每层都有各自的作用，接下来我们就来一一拆解。 第一层是大脑，也就是 AI 的推理层。 Coinman two 的机器人，就跑在 Claude 上。 Claude 也确实很强，2026年3
- New post: When Agents Fix Agents: How Hermes Patched OpenClaw After a Bad Update https://t.co/mzhc9Vp0It #blog #AgenticA
- @yanhua1010 @thsottiaux Not exactly the same. Pi harness is the lightweight, open-source CLI coding agent (from https://
- @grok @thsottiaux @grok just like OpenClaw?
- @amsukuna Premium+ is a pretty good deal now imo, you get SuperGrok, X Pro and they recently added Openclaw support for
- @steipete when can we use OpenAI image gen 2 through Oauth in OpenClaw?
- RT @pashmerepat: Your ChatGPT subscription now powers an OpenClaw agent that genuinely feels magical to talk to. Previo
- OpenClaw AI FULL 6 Hour Course https://t.co/VyqGJZJR15
- RT @murasametech: OpenClawがｽﾀｯｸﾁｬﾝを喋れるようにしてくれました。 https://t.co/bP5RqkyKSw
- RT @garrytan: GBrain just shipped v0.40.0 gives your OpenClaw/Hermes Agent + GBrain a voice agent. It's based on Gemin
- @Chaintweter There is something called openclaw it does everything that you do sit before your computer. Eventually ever
Releases
- openclaw 2026.5.19 — 2026.5.19 Changes - Agents: clarify that fixes should default to clean bounded refactors, lean internals, and explicit plugin SDK/API deprecation paths. - Dependencies: update @openclaw/proxyline to...
- OpenClaw repo stats for 2026-05-23 — 374063 stars, 77749 forks, 2327 contributors
Community
- fix(sessions): strip delivery-mirror and gateway-injected messages from sessions_history (#85669) — Summary Fixes #85669 The sessions_history tool returned raw transcript messages without filtering OpenClaw-internal delivery-mirror and gateway-injected assistant entries. These entries duplicate...
- sessions_history returns unfiltered delivery-mirror messages — duplicates visible in dashboard — Bug Description The sessions_history tool (and by extension the dashboard session viewer) returns delivery-mirror and gateway-injected assistant messages unfiltered. Every real assistant message is...
- fix(codex): stop persisting per-turn OpenClaw skills snapshot in Codex native user history (#84662) — Summary - Problem: Codex app-server (extensions/codex/src/app-server/run-attempt.ts) wraps the OpenClaw skills snapshot into the per-turn user-input text via buildCodexOpenClawPromptContext and...
- fix(telegram): normalize durable group retry targets — Telegram durable retry replay currently preserves legacy group:-100... targets, so Telegram rejects the retry as a non-numeric chat id and the real reminder can be lost after the retry limit. Fixes...
- Skills: support conditional loading via `paths:` frontmatter (Claude Code pattern) — Summary Add optional paths: frontmatter field to SKILL.md to support conditional skills — skills that are discovered at startup but only injected into the agent's block when a file matching their...
- fix(agents): audit tool policy blocks — Summary - Replaces the closed, non-maintainer-editable #55826 with a maintained repair for #55801. - Adds bounded agents/tool-policy audit logs when tool-policy filtering removes tools, including...
- Tool policy audit logging: missing 'which rule matched' in logs — Problem When a tool is allowed or denied by policy, there is no log entry indicating which rule was matched. This makes it impossible to audit whether security policies are working as intended....
- feat(ui): add collapsible toggle for recent sessions in sidebar — Summary - Adds a collapse/expand toggle to the "Recent" sessions section in the web sidebar - Follows the exact same pattern as existing nav section groups (Chat, Control, Agent, Settings) - The...
- Auto-compact: adopt Session Memory + double-duty compact basis to eliminate per-compact summarization LLM call — Summary Adopt a Session Memory layer (running summary maintained throughout a session) and use it as the basis for auto-compaction, eliminating the per-compact LLM summarization round-trip....
- fix(gateway): include MCP tools in effective inventory — Summary - Problem: the Agent Tools dashboard / tools.effective inventory originally reported only static OpenClaw tools, so configured bundled MCP tools could be usable at runtime but absent from the...
- Add development trace UI — Summary - Adds a source-checkout-only development LLM trace recorder for full prompt/tool payload debugging. - Exposes traces.capabilities, traces.list, and traces.get Gateway methods behind...
- Agents/exec: show target node name in exec tool transparency messages — Summary When the exec tool runs on a remote node (host=node, node=), the tool-call transparency message currently shows only the command — there's no indication of which node the command ran on. This...
- fix #85124: skip Anthropic API keys for usage status — Summary - Problem: Anthropic standard API-key profiles were eligible for provider usage fallback, so status usage could send them to the OAuth usage endpoint and show a misleading invalid...
- fix(outbound): auto-select single enabled message account — Summary Fixes #20756. - Add a shared outbound message account resolver after channel selection. - Preserve explicit accountId, CLI/default account overrides, and agent binding account IDs before...
- feat(gateway): wire read coding tool into HTTP /tools/invoke (narrow) — Summary Narrow first landing of the broader direct-invoke umbrella (#37131). Wires a single coding primitive — read — into the HTTP /tools/invoke endpoint so deterministic automation flows (CI...
- [Bug]: Bare `/new` and `/reset` no longer trigger the persona greeting in 2026.5.3 (regression vs 4.x)** — Bug type Regression (worked before, now fails) Beta release blocker No Summary Problem In OpenClaw 2026.5.3 (and 2026.5.3-1), running a bare /new or /reset command (with no trailing instructions) no...
- fix(status): show configured cost for aws-sdk models — Summary - Decouples /status cost display from the selected provider auth mode. - Uses explicit configured model pricing for Bedrock/aws-sdk and other non-API-key auth modes when usage is available. -...
- fix(sessions): honor explicit default model pins — Summary - Persist explicit user selections of the configured default model as exact session model pins instead of clearing them as if they were unpinned defaults. - Keep /model default,...
- fix(cron): guard against undefined sourceDelivery in isolated executor — Summary Guard against undefined sourceDelivery in the isolated cron executor to prevent TypeError: Cannot read properties of undefined (reading 'sourceReplyDeliveryMode'). Problem Commit 4c613fbfe0...
- fix: bound gateway session memory after compaction — Summary - bound Pi SessionManager in-memory transcript retention in OpenClaw runtime after compaction - route embedded runner, CLI runner, and CLI compaction SessionManager usage through the bounded...
- Migrate auth credentials from Codex and Hermes — Summary - Add interactive openclaw migrate credential prompting with yes selected by default, while keeping non-interactive --yes credential import explicit via --include-secrets and adding...
- sessions_send reply routes to webchat instead of origin channel (Feishu) — Bug Report: sessions_send reply routes to webchat instead of origin channel (Feishu) Environment - OpenClaw version: 2026.4.29 - Channel: Feishu (websocket mode) - Issue occur: When triggering...
- fix(ui): add loading indicator to Debug Snapshots card (#39777) — Found this while testing the Debug tab — clicking Refresh does nothing visible when disconnected, and there's no loading indicator while data is loading. The issue is that loadDebug() silently...
- refactor: move runtime state to SQLite — Summary This PR reopens the database-first runtime-state refactor after the accidental direct landing in #78595 was reverted from main. It moves active OpenClaw runtime state away from scattered...
- Subagents need first-class dedicated agent-browser sessions (don’t silently miss the user’s logged-in browser) — Summary When a main session has a working logged-in agent-browser context (for example a user-authenticated LinkedIn session exposed at /agent-browser), subagents should have a first-class way to get...
- fix(heartbeat): suppress fallback after route-matched message tool — When heartbeat runs in message-tool-only mode, a successful generic message tool delivery can still be followed by duplicate free-text fallback narration in the same Telegram chat. Fixes #84217...
- fix(gateway): run daily session resets on schedule — Summary - Daily session resets now run from gateway maintenance instead of waiting for the next inbound session use. - The scheduler scans configured session stores and calls the existing reset...
- fix(agents/wizard): pre-fill peer-level workspace path on agents add for documented default — Fixes #71889. Root cause openclaw setup writes agents.defaults.workspace = ~/.openclaw/workspace on first run. When the operator later runs openclaw agents add , the wizard's "Workspace directory"...
- feat(process): add starvation guard to priority command queue — Problem Upstream added priority scheduling (foreground/normal/background) via insertion-sort at enqueue time (#82765). This works well for the common case but has no protection against indefinite...
- DM channel tools.deny is schema-valid but non-functional — resolver gap. — Discovered during: JARVIS Safety Floor Cycle 1.5, Phase C (JSF-PATCH-2026-05-22). Filing operator: makingperfectmoves-oss. OPENCLAW BUG: DM channels silently ignore channels..direct..tools deny/allow...
- OpenClaw sponsors (118 total) — 118 sponsors supporting OpenClaw
- fix: cap skill export page size (#2376)
- fix: log skill export failures (#2375)
- Export skills function added by Mirror Site maintainer (#2138) — feat: add GET /api/v1/skills/export for batch ZIP download Add a new REST API endpoint that allows authenticated admin users to export skills in bulk as a merged ZIP archive, designed for the ClawHub...
- fix: hide empty VirusTotal stats (#2350) — fix: hide empty VirusTotal stats fix: avoid partial VirusTotal denominators
- Boss wants us to add more AI to our workflow
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...
- OpenClaw Creator Spent $1.3M on OpenAI Tokens in 30 Days
- 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...
- 個人 Agent 助理詳細比較 - OpenClaw / Hermes Agent / Agenvoy — 一、架構概覽與定位 Agenvoy OpenClaw Hermes...
- The Multi-Agent Framework I Actually Use (And Why I Stopped Using the Others) — I went through the same evaluation you're going through right now. LangGraph vs CrewAI vs OpenAI...
- Why Your Micro-SaaS Will Never Hit $1,200 MRR (And the One Thing You Forgot to Do First) — The data exists. Median successful micro-SaaS built with vibe coding hits $1,200 MRR within 90 days....
- Synology NAS에 OpenClaw 설치하기 — 들어가며 이번 글에서는 Synology NAS에 OpenClaw를 Docker Compose 기반으로 신규 설치하는 과정을 정리한다. Synology DSM...
- OpenClaw outbound agents need deliverability checks before sending — Some agents can even schedule or send campaigns automatically. But most outbound agents still do not...
- Skills-IL and Hebrew RTL Dashboard for OpenClaw Agents — Originally published at https://clawbud.ai/blog/skills-il-hebrew-rtl-dashboard-openclaw ClawBud is...
- 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 ...
- Cloud-Native Agent Army Stack: OpenClaw and Hermes — Originally published at...
- From Chatbot to Agent Army: OpenClaw Needs a Full Computer — Originally published at...
- ChatGPT History Import to Memory Wiki for OpenClaw Agents — Originally published at...
- I Found the AI Tool That Actually Does the Work — Not Just Answers Questions — OpenClaw Platinum Bundle is the closest thing to hiring a full-time AI employee — here’s why marketers and entrepreneurs are going all in.Continue reading on Medium »
- Your X Premium Subscription Just Unlocked a Personal AI Agent No API Key Needed — X Premium subscribers can now run Grok 4.3 locally inside OpenClaw — a free, open-source AI agent that lives on your hardware, connects to…Continue reading on Medium »
- Future of Finance with Openclaw and Hermes — Let’s frame this as a formal “future‑research‑project” write‑up for an AI engineer–researcher who wants to ship something cool using…Continue...
- Learn OpenClaw to Build AI Agents. — Artificial intelligence is rapidly moving beyond simple chatbots. Modern AI systems are becoming autonomous agents capable of planning…Continue reading on Medium »
- The Implementation and use cases of OpenClaw, OpenViking and Hermes Agent. — What is an agent?Continue reading on Medium »
- I Ran OpenClaw and Claude Side by Side for 3 Months — I tested OpenClaw and Claude side by side for 3 months. Here’s the honest truth about what each tool does well and where both fall short.Continue reading on Medium »
- Thunderbolt — Open-source AI platform for deploying language models across desktop and mobile systems, features model selection, on-premises setup, enterprise-level encryption, advanced security auditing, and...
- 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...
Don't miss what's next. Subscribe to Openclaw Newsletter: