OpenClaw Newsletter - 2026-06-17
OpenClaw Newsletter
Wednesday, June 17, 2026Top Stories
- openclaw 2026.6.8 — 2026.6.8 Highlights - Richer channel delivery: Telegram and WhatsApp are less brittle: Telegram renders structured text with tables, lists, expandable blockquotes, preserved intentional line breaks,...
- feat(migrations): add Hermes import path — Summary - add a generic migration framework and fresh-setup guard - add the built-in Hermes importer with memory/config/plugin/MCP/skill mappings - add openclaw migrate and onboarding import flags -...
- fix(agents): count message-tool source reply as user-facing reply for tool error warnings — Summary - Problem: buildEmbeddedRunPayloads has two gating variables for deliveredSourceReplyViaMessageTool that drifted apart — suppressAssistantArtifacts includes it but hasUserFacingAssistantReply...
- Containerized/self-hosted deploys can't use externalized channel plugins: openKeyedStore is gated to trusted plugins, with no supported way to trust a self-hosted channel — Summary Since the plugin unbundling (~2026.5.1), channels like @openclaw/msteams ship as separate npm packages. In a self-hosted container we deliver them by baking them into the image. Providers...
- fix(channels): resolve native /think menu levels via runtime catalog for live-discovered models — Summary - Problem: Issue #93835 reports that on Telegram, after selecting a live-discovered Ollama reasoning model (e.g. ollama/glm-5.2:cloud, whose /api/show reports capabilities: ["thinking",...
- fix(sandbox): skip read-only skill mounts when user binds target the same container path — Summary - Problem: openclaw.json sandbox.docker.binds entries targeting reserved container paths (e.g. /workspace/.agents/skills) cause the Docker daemon to reject container creation with Duplicate...
- [AI] fix(feishu): guard partial channelRuntime in monitor startup — Summary Guard partial channelRuntime in Feishu monitor startup to prevent silent incoming-message loss. - Problem: monitor.account.ts uses ?? to select channelRuntime, which picks a partial...
- fix(tui): accept both ^H (0x08) and ^? (0x7f) as valid backspace inputs — Summary - Fixes #92777: backspace key stops working in TUI on WSL2/Ubuntu because the Ubuntu WSL terminal sends ^? (0x7f) but the TUI only handled ^H (0x08). - Adds explicit check for \x7f in...
- fix(heartbeat): keep responses private when message_tool_only is active but tool not called — Summary When messages.visibleReplies: "message_tool" is configured and a heartbeat cron/wake runs, if the model produces a text reply without calling heartbeat_respond, the response was falling...
- fix: restore main ci checks — Summary - align plugin SDK surface budgets with the current checked-in public surface - drop the web-search cleanup that current main already contains Real behavior proof - Behavior or issue...
- fix(plugins): enhance plugin API compatibility diagnostic messages with troubleshooting guidance — Summary Plugin API compatibility discovery-skip warnings now include actionable troubleshooting guidance so users can self-diagnose version mismatches instead of seeing a bare error. - Problem: When...
- docs(gateway): document host agent runtime boundary policy — Summary Host-level coding-agent CLIs (Codex, host-shell) are temporary rescue tools only. Normal agent work must run inside the managed Gateway runtime. This documents the operational policy. Changes...
- [Bug] Docker sandbox fails to start with 'Duplicate mount point' when custom binds target reserved paths — Description When a user defines a custom bind mount in openclaw.json that targets the same container path as the internal hardcoded read-only skill mounts (specifically /workspace/.agents/skills),...
- fix(plugins): stop tool-discovery loads from clearing active providers — Summary - Tool-discovery plugin loads run with activate: false and a tool-only plugin scope, so the registry they build omits provider-only plugins. - installStandaloneRegistry unconditionally called...
- fix(agents): isolate RTL lines on outbound assistant text — Cherry-picked from https://github.com/openclaw/openclaw/pull/87560 Summary fix(agents): isolate RTL lines on outbound assistant text Original PR: #87560 by upstream contributor Original description:...
- fix(update): preserve local package overrides — Cherry-picked from https://github.com/openclaw/openclaw/pull/89985 Summary fix(update): preserve local package overrides Original PR: #89985 by upstream contributor Original description: Summary -...
- [Bug] Fix: Sandbox materialized skills fail to mount when Gateway runs in a Docker container (DooD) — Problem When the OpenClaw Gateway runs inside a Docker container (Docker-outside-of-Docker / DooD architecture), the gateway attempts to mount resources (materialized built-in skills, config files,...
- ci(release): harden release controls — Summary - Make release validation fail closed on missing performance evidence, missing stable soak evidence, and mutable release-evidence assets. - Add stable-main closeout verification for the...
- fix(agents): classify generic "LLM request failed." as transient time… — …out so cron fallbacks engage The generic assistant error text "LLM request failed." (GENERIC_ASSISTANT_ERROR_TEXT) is produced by formatUserFacingAssistantErrorText when the underlying provider...
- perf: lift realpath cache scope in buildInstalledManifestRegistryIndexKey — > AI-assisted PR. Implemented and verified with Claude (Claude Code). I have reviewed and understand the change and own the final diff. Summary buildInstalledManifestRegistryIndexKey() creates a new...
- Fix sessions.list for literal per-agent store paths — Cherry-picked from https://github.com/openclaw/openclaw/pull/54716 Summary Fix sessions.list for literal per-agent store paths Original PR: #54716 by upstream contributor Original description:...
- fix(sandbox): skip custom binds colliding with protected skill mount destinations — Summary - What: Docker sandbox container creation fails with Duplicate mount point when a user-defined custom bind targets a container path that overlaps with internal read-only skill mounts (e.g....
- fix(CLI) Markdown table columns misaligned with CJK characters on Telegram and Discord — Cherry-picked from https://github.com/openclaw/openclaw/pull/55596 Summary fix(CLI) Markdown table columns misaligned with CJK characters on Telegram and Discord Original PR: #55596 by upstream...
- feat(onboard): streamline setup with agent-assisted configuration — Summary - Make bare openclaw and canonical openclaw onboard choose the shortest runnable local path, while keeping openclaw setup as the documented onboarding alias. - Offer migration from supported...
- feat: add devcontainer for cross-platform development — Cherry-picked from https://github.com/openclaw/openclaw/pull/62503 Summary feat: add devcontainer for cross-platform development Original PR: #62503 by upstream contributor Original description:...
- perf(agents): cache provider tool schema normalization — Summary - Cache provider-normalized tool schema parameters for repeated embedded agent turns, but only when OpenClaw can identify one of its bundled SDK tool-compat normalizers. - Keep the cache...
- fix(config): allow historyLimit: 0 in GroupChatSchema — Cherry-picked from https://github.com/openclaw/openclaw/pull/65359 Summary fix(config): allow historyLimit: 0 in GroupChatSchema Original PR: #65359 by upstream contributor Original description:...
- fix: strip volatile output from exec result hash to fix no-progress detection (#93917) — Summary The genericRepeat detector in tool-loop-detection.ts has two detection paths: 1. Warning path: uses recentCount — counts all calls with identical args, regardless of result. ✅ Works...
- fix(mattermost): backfill thread history from server when in-memory window is empty — Summary Fixes #93204 — Mattermost thread context is lost after gateway restart or session clear because the in-memory channelHistories map has no server-side backfill. When a user @mentions the bot...
- fix: support baseURL alias in memory embedding provider config resolution — Summary - When users configure models.providers.openai with baseURL (camelCase) pointing to a local Ollama or OpenAI-compatible endpoint, the memory embedding system ignored the baseURL field and...
- OpenClaw repo stats for 2026-06-17 — 379126 stars, 79349 forks, 2601 contributors
- OpenClaw sponsors (109 total) — 109 sponsors supporting OpenClaw
- fix: add install sort support to package catalog APIs — - Add install sorting to package and plugin catalog API paths. - Reject removed downloads sort requests with 400. - Keep recommended browse stable during recommendation-score backfill. - Normalize...
- build(deps-dev): bump the development-minor-and-patch group across 1 directory with 6 updates (#2689) — Bumps the development-minor-and-patch group with 6 updates in the / directory: Updates @playwright/test from 1.60.0 to 1.61.0 - Release notes - Commits Updates @vitest/coverage-v8 from 4.1.8 to 4.1.9...
- fix: clarify publisher abuse ban emails (#2697)
- docs: fix acceptable usage table formatting
- docs: clarify acceptable usage policy
- fix: remove retired capability metadata surfaces — Remove retired capabilities/capabilityTags/executesCode surfaces for skills and plugins, add guarded cleanup tooling for legacy rows, and sanitize pre-cleanup legacy seed/release payloads.
- fix: seed worktree dev previews locally (#2693)
- docs: strengthen ClawHub PR maintainer reviews (#2692)
- Remove retired moderation surfaces (#2658) — chore: remove retired moderation surfaces test: cover retired moderation action wrappers chore: retain moderation security gates test: cover retired moderation compatibility paths fix: stop settings...
- refactor: remove root install telemetry schema (#2682)
- Rank skills by installs and harden install telemetry — docs: add install telemetry ranking spec feat: rank skills by installs and dedupe install telemetry test: cover install telemetry dedupe cleanup fix: clean install dedupes on skill hard delete fix:...
- fix: remove registry backup index writes
- update coderabbit
- openclaw v2026.6.8 on npm — Multi-channel AI gateway with extensible messaging integrations
- openclaw-cli 2026.6.6 on Homebrew — Your own personal AI assistant
- 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...
- You are not supposed to install OpenClaw on your personal computer
- StepFun 3.5 Flash is #1 cost-effective model for OpenClaw tasks (300 battles)
- Show HN: Klaus – OpenClaw on a VM, batteries included
- I thought the cheap model would save my OpenClaw bill, then I watched $100 disappear in 2 days — A practical DEV post for OpenClaw users: why using one cheap model everywhere often increases total cost, how to route by task risk instead, and why c
- Building a consent-preserving auth.md skill for OpenClaw agents — Signup forms were designed for humans with browsers. AI agents get a 401 and have nowhere to...
- OpenClaw Deep Dive: เมื่อ Coding Agent อยากเป็นมากกว่าแค่คนเขียน Code — ใน Session “OpenClaw Deep Dive: เมื่อ Coding Agent...
- Looking for an OpenClaw Alternative? Here Are 7 Worth Exploring in 2026 — AI agents have evolved far beyond simple chatbots. Tools like OpenClaw can automate tasks, remember context, interact with apps, and…Continue reading on Medium »
- OpenClaw kind of became my personal life-admin assistant and I didn’t expect that — I started using OpenClaw just to mess around, but over the last couple of months it’s turned into the thing that keeps my personal life…Continue reading on Medium »
- OpenClaw Cost Guide 2026: Pricing, Hosting & How to Save 90% on API Fees — Quick Answer: What Will OpenClaw Actually Cost You?Continue reading on Medium »
- OpenClaw for Cybersecurity: Powerful… If You Use It Correctly — How to safely automate workflows without connecting to sensitive systemsContinue reading on Medium »
- I Spent Six Months Building What WhatsApp Just Shipped for Free — Why AI agents will supercharge the tools we already use — not replace them.Continue reading on Medium »
- Prototipo de seguridad con agente de IA + ML — No todas las tecnologías evolucionan al mismo ritmo.Continue reading on Medium »
- The One Where You Get Hired at Central Perk — Congratulations: you’ve just been hired at a virtual coffee shop — staffed entirely by software, run by a sitcom cast.Continue reading on Medium »
- Telegram Ban in India Until 22 June: How I Fixed OpenClaw and Telegram AI Agents Running on Indian… — If you woke up one morning and found that your OpenClaw instance had suddenly stopped responding on Telegram, you’re not alone.Continue reading on Medium »
- Cost-Aware Agentic AI Is Harder Than Picking a Cheaper Model — A practical OpenClaw experiment in decomposition, routing, context control, and execution evidenceContinue reading on Medium »
- NanoClaw — Lightweight AI assistant runs in Apple containers with isolated Linux container processes for superior OS-level security, transparency, and privacy compared to complex chatbots, enabling quick code...
- Maid (AI) — Maid is a cross-platform Flutter app for interfacing with GGUF / llama.cpp models locally, and with Ollama and OpenAI models remotely.
- openclaw been right on MUSHU twice already. trusting the scan ┃🧫┃ $MUSHU ┃💰┃ MC : $449.9K ┃🎯┃ TARGET MC : $1.2M ┃📈┃
- just checked POKE —openclaw showing strong upside signals ┃🧫┃ $POKE ┃💰┃ MC : $26.9K ┃🎯┃ TARGET MC : $382.9K ┃📈┃ POT
- just ran MITCH through openclaw. risk is low, upside is real ┃🧫┃ $MITCH ┃💰┃ MC : $378.6K ┃🎯┃ TARGET MC : $1.0M ┃📈┃
- EMBER signals are clean on openclaw. this one's on my radar 🦞 ┃🧫┃ $EMBER ┃💰┃ MC : $25.9K ┃🎯┃ TARGET MC : $406.6K ┃📈
- openclaw on Gnome —signals all pointing up ┃🧫┃ $Gnome ┃💰┃ MC : $350.5K ┃🎯┃ TARGET MC : $1.3M ┃📈┃ POTENTIAL : +275%
- openclaw flagged CLAW early. price is moving exactly as predicted ┃🧫┃ $CLAW ┃💰┃ MC : $3.0M ┃🎯┃ TARGET MC : $9.8M ┃📈
- openclaw flagged bluechip as a strong hold. not surprised ┃🧫┃ $bluechip ┃💰┃ MC : $107.3K ┃🎯┃ TARGET MC : $575.0K ┃📈
- openclaw called GAPLA bullish. i'm listening ┃🧫┃ $GAPLA ┃💰┃ MC : $85.8K ┃🎯┃ TARGET MC : $527.8K ┃📈┃ POTENTIAL : +51
- @github OpenClaw complicates this further. 345K+ GitHub stars, connects to 50+ messaging apps, runs community-built skil
- fish is looking good on openclaw right now ngl ┃🧫┃ $fish ┃💰┃ MC : $131.8K ┃🎯┃ TARGET MC : $888.6K ┃📈┃ POTENTIAL : +
- CODEC momentum score just came in hot on openclaw 🔥 ┃🧫┃ $CODEC ┃💰┃ MC : $3.7M ┃🎯┃ TARGET MC : $13.1M ┃📈┃ POTENTIAL
- openclaw called ATTENTION bullish. i'm listening ┃🧫┃ $ATTENTION ┃💰┃ MC : $713.2K ┃🎯┃ TARGET MC : $2.5M ┃📈┃ POTENTIA
- openclaw flagged DooDoo as a strong hold. not surprised ┃🧫┃ $DooDoo ┃💰┃ MC : $19.8K ┃🎯┃ TARGET MC : $237.6K ┃📈┃ POT
- RETARD is looking good on openclaw right now ngl ┃🧫┃ $RETARD ┃💰┃ MC : $85.4K ┃🎯┃ TARGET MC : $323.0K ┃📈┃ POTENTIAL
- just checked CLOUT on openclaw—forecast looking clean 🦞 ┃🧫┃ $CLOUT ┃💰┃ MC : $16.5K ┃🎯┃ TARGET MC : $199.7K ┃📈┃ POTE
- openclaw on POKECR —momentum strong, risk low. easy ┃🧫┃ $POKECR ┃💰┃ MC : $15.0K ┃🎯┃ TARGET MC : $215.1K ┃📈┃ POTENTI
- openclaw called it on Fine again. the model doesn't miss ┃🧫┃ $Fine ┃💰┃ MC : $54.4K ┃🎯┃ TARGET MC : $357.7K ┃📈┃ POTE
- just checked Wasabi on openclaw—forecast looking clean 🦞 ┃🧫┃ $Wasabi ┃💰┃ MC : $1.1M ┃🎯┃ TARGET MC : $3.8M ┃📈┃ POTEN
- openclaw on ELGNT —signals all pointing up ┃🧫┃ $ELGNT ┃💰┃ MC : $15.0K ┃🎯┃ TARGET MC : $230.8K ┃📈┃ POTENTIAL : +1443
- just ran MIRARI through openclaw. risk is low, upside is real ┃🧫┃ $MIRARI ┃💰┃ MC : $10.2K ┃🎯┃ TARGET MC : $140.1K ┃
Trending on X
- openclaw been right on MUSHU twice already. trusting the scan ┃🧫┃ $MUSHU ┃💰┃ MC : $449.9K ┃🎯┃ TARGET MC : $1.2M ┃📈┃
- just checked POKE —openclaw showing strong upside signals ┃🧫┃ $POKE ┃💰┃ MC : $26.9K ┃🎯┃ TARGET MC : $382.9K ┃📈┃ POT
- just ran MITCH through openclaw. risk is low, upside is real ┃🧫┃ $MITCH ┃💰┃ MC : $378.6K ┃🎯┃ TARGET MC : $1.0M ┃📈┃
- EMBER signals are clean on openclaw. this one's on my radar 🦞 ┃🧫┃ $EMBER ┃💰┃ MC : $25.9K ┃🎯┃ TARGET MC : $406.6K ┃📈
- openclaw on Gnome —signals all pointing up ┃🧫┃ $Gnome ┃💰┃ MC : $350.5K ┃🎯┃ TARGET MC : $1.3M ┃📈┃ POTENTIAL : +275%
- openclaw flagged CLAW early. price is moving exactly as predicted ┃🧫┃ $CLAW ┃💰┃ MC : $3.0M ┃🎯┃ TARGET MC : $9.8M ┃📈
- openclaw flagged bluechip as a strong hold. not surprised ┃🧫┃ $bluechip ┃💰┃ MC : $107.3K ┃🎯┃ TARGET MC : $575.0K ┃📈
- openclaw called GAPLA bullish. i'm listening ┃🧫┃ $GAPLA ┃💰┃ MC : $85.8K ┃🎯┃ TARGET MC : $527.8K ┃📈┃ POTENTIAL : +51
- @github OpenClaw complicates this further. 345K+ GitHub stars, connects to 50+ messaging apps, runs community-built skil
- fish is looking good on openclaw right now ngl ┃🧫┃ $fish ┃💰┃ MC : $131.8K ┃🎯┃ TARGET MC : $888.6K ┃📈┃ POTENTIAL : +
- CODEC momentum score just came in hot on openclaw 🔥 ┃🧫┃ $CODEC ┃💰┃ MC : $3.7M ┃🎯┃ TARGET MC : $13.1M ┃📈┃ POTENTIAL
- openclaw called ATTENTION bullish. i'm listening ┃🧫┃ $ATTENTION ┃💰┃ MC : $713.2K ┃🎯┃ TARGET MC : $2.5M ┃📈┃ POTENTIA
- openclaw flagged DooDoo as a strong hold. not surprised ┃🧫┃ $DooDoo ┃💰┃ MC : $19.8K ┃🎯┃ TARGET MC : $237.6K ┃📈┃ POT
- RETARD is looking good on openclaw right now ngl ┃🧫┃ $RETARD ┃💰┃ MC : $85.4K ┃🎯┃ TARGET MC : $323.0K ┃📈┃ POTENTIAL
- just checked CLOUT on openclaw—forecast looking clean 🦞 ┃🧫┃ $CLOUT ┃💰┃ MC : $16.5K ┃🎯┃ TARGET MC : $199.7K ┃📈┃ POTE
- openclaw on POKECR —momentum strong, risk low. easy ┃🧫┃ $POKECR ┃💰┃ MC : $15.0K ┃🎯┃ TARGET MC : $215.1K ┃📈┃ POTENTI
- openclaw called it on Fine again. the model doesn't miss ┃🧫┃ $Fine ┃💰┃ MC : $54.4K ┃🎯┃ TARGET MC : $357.7K ┃📈┃ POTE
- just checked Wasabi on openclaw—forecast looking clean 🦞 ┃🧫┃ $Wasabi ┃💰┃ MC : $1.1M ┃🎯┃ TARGET MC : $3.8M ┃📈┃ POTEN
- openclaw on ELGNT —signals all pointing up ┃🧫┃ $ELGNT ┃💰┃ MC : $15.0K ┃🎯┃ TARGET MC : $230.8K ┃📈┃ POTENTIAL : +1443
- just ran MIRARI through openclaw. risk is low, upside is real ┃🧫┃ $MIRARI ┃💰┃ MC : $10.2K ┃🎯┃ TARGET MC : $140.1K ┃
Releases
- openclaw 2026.6.8 — 2026.6.8 Highlights - Richer channel delivery: Telegram and WhatsApp are less brittle: Telegram renders structured text with tables, lists, expandable blockquotes, preserved intentional line breaks,...
- OpenClaw repo stats for 2026-06-17 — 379126 stars, 79349 forks, 2601 contributors
- openclaw v2026.6.8 on npm — Multi-channel AI gateway with extensible messaging integrations
- openclaw-cli 2026.6.6 on Homebrew — Your own personal AI assistant
Community
- feat(migrations): add Hermes import path — Summary - add a generic migration framework and fresh-setup guard - add the built-in Hermes importer with memory/config/plugin/MCP/skill mappings - add openclaw migrate and onboarding import flags -...
- fix(agents): count message-tool source reply as user-facing reply for tool error warnings — Summary - Problem: buildEmbeddedRunPayloads has two gating variables for deliveredSourceReplyViaMessageTool that drifted apart — suppressAssistantArtifacts includes it but hasUserFacingAssistantReply...
- Containerized/self-hosted deploys can't use externalized channel plugins: openKeyedStore is gated to trusted plugins, with no supported way to trust a self-hosted channel — Summary Since the plugin unbundling (~2026.5.1), channels like @openclaw/msteams ship as separate npm packages. In a self-hosted container we deliver them by baking them into the image. Providers...
- fix(channels): resolve native /think menu levels via runtime catalog for live-discovered models — Summary - Problem: Issue #93835 reports that on Telegram, after selecting a live-discovered Ollama reasoning model (e.g. ollama/glm-5.2:cloud, whose /api/show reports capabilities: ["thinking",...
- fix(sandbox): skip read-only skill mounts when user binds target the same container path — Summary - Problem: openclaw.json sandbox.docker.binds entries targeting reserved container paths (e.g. /workspace/.agents/skills) cause the Docker daemon to reject container creation with Duplicate...
- [AI] fix(feishu): guard partial channelRuntime in monitor startup — Summary Guard partial channelRuntime in Feishu monitor startup to prevent silent incoming-message loss. - Problem: monitor.account.ts uses ?? to select channelRuntime, which picks a partial...
- fix(tui): accept both ^H (0x08) and ^? (0x7f) as valid backspace inputs — Summary - Fixes #92777: backspace key stops working in TUI on WSL2/Ubuntu because the Ubuntu WSL terminal sends ^? (0x7f) but the TUI only handled ^H (0x08). - Adds explicit check for \x7f in...
- fix(heartbeat): keep responses private when message_tool_only is active but tool not called — Summary When messages.visibleReplies: "message_tool" is configured and a heartbeat cron/wake runs, if the model produces a text reply without calling heartbeat_respond, the response was falling...
- fix: restore main ci checks — Summary - align plugin SDK surface budgets with the current checked-in public surface - drop the web-search cleanup that current main already contains Real behavior proof - Behavior or issue...
- fix(plugins): enhance plugin API compatibility diagnostic messages with troubleshooting guidance — Summary Plugin API compatibility discovery-skip warnings now include actionable troubleshooting guidance so users can self-diagnose version mismatches instead of seeing a bare error. - Problem: When...
- docs(gateway): document host agent runtime boundary policy — Summary Host-level coding-agent CLIs (Codex, host-shell) are temporary rescue tools only. Normal agent work must run inside the managed Gateway runtime. This documents the operational policy. Changes...
- [Bug] Docker sandbox fails to start with 'Duplicate mount point' when custom binds target reserved paths — Description When a user defines a custom bind mount in openclaw.json that targets the same container path as the internal hardcoded read-only skill mounts (specifically /workspace/.agents/skills),...
- fix(plugins): stop tool-discovery loads from clearing active providers — Summary - Tool-discovery plugin loads run with activate: false and a tool-only plugin scope, so the registry they build omits provider-only plugins. - installStandaloneRegistry unconditionally called...
- fix(agents): isolate RTL lines on outbound assistant text — Cherry-picked from https://github.com/openclaw/openclaw/pull/87560 Summary fix(agents): isolate RTL lines on outbound assistant text Original PR: #87560 by upstream contributor Original description:...
- fix(update): preserve local package overrides — Cherry-picked from https://github.com/openclaw/openclaw/pull/89985 Summary fix(update): preserve local package overrides Original PR: #89985 by upstream contributor Original description: Summary -...
- [Bug] Fix: Sandbox materialized skills fail to mount when Gateway runs in a Docker container (DooD) — Problem When the OpenClaw Gateway runs inside a Docker container (Docker-outside-of-Docker / DooD architecture), the gateway attempts to mount resources (materialized built-in skills, config files,...
- ci(release): harden release controls — Summary - Make release validation fail closed on missing performance evidence, missing stable soak evidence, and mutable release-evidence assets. - Add stable-main closeout verification for the...
- fix(agents): classify generic "LLM request failed." as transient time… — …out so cron fallbacks engage The generic assistant error text "LLM request failed." (GENERIC_ASSISTANT_ERROR_TEXT) is produced by formatUserFacingAssistantErrorText when the underlying provider...
- perf: lift realpath cache scope in buildInstalledManifestRegistryIndexKey — > AI-assisted PR. Implemented and verified with Claude (Claude Code). I have reviewed and understand the change and own the final diff. Summary buildInstalledManifestRegistryIndexKey() creates a new...
- Fix sessions.list for literal per-agent store paths — Cherry-picked from https://github.com/openclaw/openclaw/pull/54716 Summary Fix sessions.list for literal per-agent store paths Original PR: #54716 by upstream contributor Original description:...
- fix(sandbox): skip custom binds colliding with protected skill mount destinations — Summary - What: Docker sandbox container creation fails with Duplicate mount point when a user-defined custom bind targets a container path that overlaps with internal read-only skill mounts (e.g....
- fix(CLI) Markdown table columns misaligned with CJK characters on Telegram and Discord — Cherry-picked from https://github.com/openclaw/openclaw/pull/55596 Summary fix(CLI) Markdown table columns misaligned with CJK characters on Telegram and Discord Original PR: #55596 by upstream...
- feat(onboard): streamline setup with agent-assisted configuration — Summary - Make bare openclaw and canonical openclaw onboard choose the shortest runnable local path, while keeping openclaw setup as the documented onboarding alias. - Offer migration from supported...
- feat: add devcontainer for cross-platform development — Cherry-picked from https://github.com/openclaw/openclaw/pull/62503 Summary feat: add devcontainer for cross-platform development Original PR: #62503 by upstream contributor Original description:...
- perf(agents): cache provider tool schema normalization — Summary - Cache provider-normalized tool schema parameters for repeated embedded agent turns, but only when OpenClaw can identify one of its bundled SDK tool-compat normalizers. - Keep the cache...
- fix(config): allow historyLimit: 0 in GroupChatSchema — Cherry-picked from https://github.com/openclaw/openclaw/pull/65359 Summary fix(config): allow historyLimit: 0 in GroupChatSchema Original PR: #65359 by upstream contributor Original description:...
- fix: strip volatile output from exec result hash to fix no-progress detection (#93917) — Summary The genericRepeat detector in tool-loop-detection.ts has two detection paths: 1. Warning path: uses recentCount — counts all calls with identical args, regardless of result. ✅ Works...
- fix(mattermost): backfill thread history from server when in-memory window is empty — Summary Fixes #93204 — Mattermost thread context is lost after gateway restart or session clear because the in-memory channelHistories map has no server-side backfill. When a user @mentions the bot...
- fix: support baseURL alias in memory embedding provider config resolution — Summary - When users configure models.providers.openai with baseURL (camelCase) pointing to a local Ollama or OpenAI-compatible endpoint, the memory embedding system ignored the baseURL field and...
- OpenClaw sponsors (109 total) — 109 sponsors supporting OpenClaw
- fix: add install sort support to package catalog APIs — - Add install sorting to package and plugin catalog API paths. - Reject removed downloads sort requests with 400. - Keep recommended browse stable during recommendation-score backfill. - Normalize...
- build(deps-dev): bump the development-minor-and-patch group across 1 directory with 6 updates (#2689) — Bumps the development-minor-and-patch group with 6 updates in the / directory: Updates @playwright/test from 1.60.0 to 1.61.0 - Release notes - Commits Updates @vitest/coverage-v8 from 4.1.8 to 4.1.9...
- fix: clarify publisher abuse ban emails (#2697)
- docs: fix acceptable usage table formatting
- docs: clarify acceptable usage policy
- fix: remove retired capability metadata surfaces — Remove retired capabilities/capabilityTags/executesCode surfaces for skills and plugins, add guarded cleanup tooling for legacy rows, and sanitize pre-cleanup legacy seed/release payloads.
- fix: seed worktree dev previews locally (#2693)
- docs: strengthen ClawHub PR maintainer reviews (#2692)
- Remove retired moderation surfaces (#2658) — chore: remove retired moderation surfaces test: cover retired moderation action wrappers chore: retain moderation security gates test: cover retired moderation compatibility paths fix: stop settings...
- refactor: remove root install telemetry schema (#2682)
- Rank skills by installs and harden install telemetry — docs: add install telemetry ranking spec feat: rank skills by installs and dedupe install telemetry test: cover install telemetry dedupe cleanup fix: clean install dedupes on skill hard delete fix:...
- fix: remove registry backup index writes
- update coderabbit
News
- 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...
- You are not supposed to install OpenClaw on your personal computer
- StepFun 3.5 Flash is #1 cost-effective model for OpenClaw tasks (300 battles)
- Show HN: Klaus – OpenClaw on a VM, batteries included
- I thought the cheap model would save my OpenClaw bill, then I watched $100 disappear in 2 days — A practical DEV post for OpenClaw users: why using one cheap model everywhere often increases total cost, how to route by task risk instead, and why c
- Building a consent-preserving auth.md skill for OpenClaw agents — Signup forms were designed for humans with browsers. AI agents get a 401 and have nowhere to...
- OpenClaw Deep Dive: เมื่อ Coding Agent อยากเป็นมากกว่าแค่คนเขียน Code — ใน Session “OpenClaw Deep Dive: เมื่อ Coding Agent...
- Looking for an OpenClaw Alternative? Here Are 7 Worth Exploring in 2026 — AI agents have evolved far beyond simple chatbots. Tools like OpenClaw can automate tasks, remember context, interact with apps, and…Continue reading on Medium »
- OpenClaw kind of became my personal life-admin assistant and I didn’t expect that — I started using OpenClaw just to mess around, but over the last couple of months it’s turned into the thing that keeps my personal life…Continue reading on Medium »
- OpenClaw Cost Guide 2026: Pricing, Hosting & How to Save 90% on API Fees — Quick Answer: What Will OpenClaw Actually Cost You?Continue reading on Medium »
- OpenClaw for Cybersecurity: Powerful… If You Use It Correctly — How to safely automate workflows without connecting to sensitive systemsContinue reading on Medium »
- I Spent Six Months Building What WhatsApp Just Shipped for Free — Why AI agents will supercharge the tools we already use — not replace them.Continue reading on Medium »
- Prototipo de seguridad con agente de IA + ML — No todas las tecnologías evolucionan al mismo ritmo.Continue reading on Medium »
- The One Where You Get Hired at Central Perk — Congratulations: you’ve just been hired at a virtual coffee shop — staffed entirely by software, run by a sitcom cast.Continue reading on Medium »
- Telegram Ban in India Until 22 June: How I Fixed OpenClaw and Telegram AI Agents Running on Indian… — If you woke up one morning and found that your OpenClaw instance had suddenly stopped responding on Telegram, you’re not alone.Continue reading on Medium »
- Cost-Aware Agentic AI Is Harder Than Picking a Cheaper Model — A practical OpenClaw experiment in decomposition, routing, context control, and execution evidenceContinue reading on Medium »
- NanoClaw — Lightweight AI assistant runs in Apple containers with isolated Linux container processes for superior OS-level security, transparency, and privacy compared to complex chatbots, enabling quick code...
- Maid (AI) — Maid is a cross-platform Flutter app for interfacing with GGUF / llama.cpp models locally, and with Ollama and OpenAI models remotely.
Don't miss what's next. Subscribe to Openclaw Newsletter: