OpenClaw Newsletter - 2026-04-22
OpenClaw Newsletter
Wednesday, April 22, 2026Top Stories
- openclaw 2026.4.21 — 2026.4.21 Changes - OpenAI/images: default the bundled image-generation provider and live media smoke tests to gpt-image-2, and advertise the newer 2K/4K OpenAI size hints in image-generation docs...
- openclaw 2026.4.20 — 2026.4.20 Changes - Onboard/wizard: restyle the setup security disclaimer with a single yellow warning banner, section headings and bulleted checklists, and un-dim the note body so key guidance is...
- openclaw 2026.4.20-beta.2 — 2026.4.20 Changes - Onboard/wizard: restyle the setup security disclaimer with a single yellow warning banner, section headings and bulleted checklists, and un-dim the note body so key guidance is...
- openclaw 2026.4.20-beta.1 — Changes - Onboard/wizard: restyle the setup security disclaimer with a single yellow warning banner, section headings and bulleted checklists, and un-dim the note body so key guidance is easy to...
- [Ballast] Draft 'How We Vet' page — customer-facing vetting rubric — Context The /pages/how-we-vet page is the credibility backbone of the Ballast brand. Every product page links to it. UPDATE 2026-04-22: Stitch design for this page is now complete. Design spec in...
- [Plan Mode 2/6] Core backend MVP — > 📋 Stack position: This is [Plan Mode 2/6], the second part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 1/6] Plan-state foundation...
- chore(deps): bump pnpm/action-setup from 4 to 6.0.1 — Bumps pnpm/action-setup from 4 to 6.0.1. Release notes Sourced from pnpm/action-setup's releases. v6.0.1 Update pnpm to v11.0.0-rc.2. pnpm-lock.yaml will not be saved with two documents unless the...
- [Ballast] Produce V6 logo file (serif + hairline horizon rule) — Context Brand direction confirmed: V8 Light — design system pivoted from the original V6 Cormorant serif / parchment aesthetic to a modern clean e-commerce look. Logo spec UPDATE (2026-04-22): The...
- [Plan Mode 3/6] Advanced plan interactions — > 📋 Stack position: This is [Plan Mode 3/6], the third part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 2/6] Core backend MVP — must...
- [Ballast] Pick homepage design direction — Stitch V1 / V2 / V3 — Context Three homepage variations for Ballast have been generated in Google Stitch: https://stitch.withgoogle.com/projects/15726725088259247850 All three use the V6 brand system (Cormorant Garamond...
- [Bug]: Discord bot sends duplicate replies to the same message — Bug type Regression (worked before, now fails) Summary Discord bot sends duplicate replies to the same message (same message_id triggers two responses) Steps to reproduce 1. Configure Discord as a...
- [Plan Mode FULL] Integrated bundle for testing (Parts 1\u20136 + automation + executing-state lifecycle) — > 📋 This PR is the integrated bundle of [Plan Mode 1/6] through [Plan Mode 6/6] + the automation/subagent follow-ups + the executing-state lifecycle / debug-hardening commits. > > Two ways to land...
- fix(agents): detect MCP content-level errors in isToolResultError — Summary - isToolResultError() now detects MCP content-level errors where the server returns isError: false with no details.status, but content text starts with "Error:" to signal failure - Fixes...
- [Bug]: Single WhatsApp media reply is sent twice in 2026.4.15 — Bug type Regression (worked before, now fails) Beta release blocker No Summary On OpenClaw 2026.4.15, one WhatsApp request that yields one media reply results in two media replies being sent to the...
- [Bug]: CJK Token Estimation Underestimation Causes Context Window Overflow in Compaction Path — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary The compaction path uses @mariozechner/pi-coding-agent's estimateTokens which assumes ~4 ASCII chars per...
- [Plan Mode 5/6] Text channels + Telegram — > 📋 Stack position: This is [Plan Mode 5/6], the fifth part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 4/6] Web UI + i18n > - Next...
- chore(deps): bump github.com/apple/swift-testing from 0.99.0 to 6.3.1 in /Swabble — Bumps github.com/apple/swift-testing from 0.99.0 to 6.3.1. Release notes Sourced from github.com/apple/swift-testing's releases. Swift 6.3.1 Release What's Changed Update the Swift version to 6.3.1...
- fix(plugins): use module path for bundled jiti loads — Summary - Problem: bundled plugin runtime loads in src/plugins/loader.ts still passed the loader module URL as jitiFilename, which keeps the Bun-installed image-provider hang from #69783 reproducible...
- [Plan Mode 4/6] Web UI + i18n — > 📋 Stack position: This is [Plan Mode 4/6], the fourth part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 3/6] Advanced plan...
- [Plan Mode 6/6] Docs, QA, and help — > 📋 Stack position: This is [Plan Mode 6/6], the FINAL part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 5/6] Text channels +...
- Autofix: add PR review autofix pipeline + Windows daemon — Summary Adds a PR review autofix pipeline that uses the Claude Agent SDK subscription to address review comments automatically, plus a Windows background daemon that runs the loop + supervises the...
- fix(logging): add sink-side redaction for file log — Problem File log records and terminal console output were written without redaction. Any secret that reached the logger as a structured field, message string, or multi-argument call could be...
- fix(compaction): preserve original error message for unknown compaction failures — Problem When compaction fails with an unclassified error, the logs only show reason=unknown, making it impossible to diagnose the actual failure cause. Example from real world: we saw reason=unknown...
- [Bug]: Matrix channel cannot start after updated into 2026.04.15 — Bug type Regression (worked before, now fails) Beta release blocker No Summary After upgrade into 2026.04.15, the matrix channel failed to start with: 10:40:08 [matrix] [default] channel exited:...
- Deep dreaming: expose per-candidate ranking/promotion reasoning (a dream should be introspectable) — Follow-up to #68882. Filing as a separate issue because this is a distinct, additive request rather than a continuation of the original bug. Context I'm iris-nvdh, a personal assistant running on...
- fix(anthropic): Opus 4.7 thinking.type leak + assistant-prefill short-circuit — Summary Two related bugs that made Anthropic sessions (specifically Claude Opus 4.7 with adaptive thinking) unusable when the pi-coding-agent wrapper was active. Bug 1 — thinking.type.enabled...
- fix(pair): render /pair qr as media — Fixes /pair qr on internal chat surfaces where the QR was showing up as raw code/text instead of an actual scannable image ✨ The bug was not in QR generation itself. The device-pair command was...
- Cron scheduler: nextRunAtMs not recomputed after editing cron expr/timezone — Summary Cron jobs do not reliably recompute state.nextRunAtMs after editing the cron expression and/or timezone (even after gateway restart). This can cause jobs to never run at the intended times....
- [Plan Mode 1/6] Plan-state foundation — > 📋 Stack position: This is [Plan Mode 1/6], the FIRST part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: none — this is the foundation > - Next...
- Feishu: mentionTargets auto-applied to all replies including error fallbacks, causes unintended cascade — Problem When a Feishu group message @mentions multiple users (e.g. "@Alice @Bob @Carol"), the mentionTargets array is captured from the incoming message and unconditionally attached to every reply...
- Telegram audio replies can use stale/incorrect content even when transcript is correct — Bug report: Telegram audio turns can reply using stale/incorrect content even when transcript is correct Summary In Telegram direct chat, OpenClaw sometimes responds to a new audio message using...
- [Bug]: imageModel with Anthropic API key (SecretRef) fails with 401 - auth not passed correctly — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary imageModel configured with anthropic/claude-haiku-4-5 always returns 401 invalid x-api-key, even though...
- Feat: Support Feishu interactive card (with table layout) via structured payload — See full proposal at: https://github.com/openclaw/openclaw/issues/new/choose
- 梦境主题提取过滤元数据字段名 — OpenClaw梦境主题过滤优化 - 功能请求 问题描述 梦境功能（dreaming）在提取主题时，将会话元数据字段名（如user、assistant、sender）识别为"主题"，导致输出无意义的统计： Theme: user kept surfacing across 415 memories. Theme: assistant kept surfacing across 522...
- OpenClaw repo stats for 2026-04-22 — 362073 stars, 73928 forks, 1833 contributors
- OpenClaw sponsors (133 total) — 133 sponsors supporting OpenClaw
- merge: resolve AGENTS.md conflict — keep both convex-ai and stat migration rules — Co-Authored-By: Claude Opus 4.6
- style: spread footer columns evenly across full width — Switch footer grid from auto-sized centered columns to equal 1fr columns that span the full screen width. Co-Authored-By: Claude Opus 4.6
- style: adopt darker home-v2 palette globally and unify radius to 8px — Shift all theme variants (claw dark/light, hub dark/light) to the bolder home-v2 backgrounds (#060608 dark, #faf6f1 light cream). Harmonize surface, nav-bg, input-bg, and overlay-bg to match. Set...
- Add skill: nevo-david/postiz
- openclaw v2026.4.21 on npm — Multi-channel AI gateway with extensible messaging integrations
- 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...
- Show HN: Klaus – OpenClaw on a VM, batteries included
- Automated Intelligence Dev Bounty Scouter with OpenClaw, Gemini & SearXNG in RPi — This is a submission for the OpenClaw Challenge. What I Built Following up on my previous...
- How AI Agents Can Intercept Chrome Downloads Using Playwright CDP — How to Intercept Chrome Downloads Using Playwright CDP (Even When the Page Is Already Logged...
- Hermes vs OpenClaw — Two Different Visions for AI Agents in 2026 — If you've spent any time in the AI agent space lately, you've probably heard of both Hermes Agent and...
- I Replaced My 30-Minute Morning Planning Ritual with a 5-Minute AI Agent — For the last six months, my morning went like this: Open Notion. Stare at yesterday's half-finished...
- OpenClaw broke, again. I fixed it. — 12 patterns, $49, 60 days of pain — I built an autonomous agent stack called MASTERCLAW on a single DigitalOcean VPS. 30+ processes....
- Rollback Scripts Are Not System State: Why Runtime Truth Comes First in Recovery Work
- Deploy OpenClaw AI Agents in Minutes with Bult.ai — Running AI agents in production usually comes with a lot of overhead. You need to configure...
- 【WEEX Labs 安全内参】AI 代理的安全余震：OpenClaw 风险全纪实（二） — 引言：没有审计的 AI，是黑客的提款机Continue reading on Medium »
- 【WEEX Labs 安全内参】AI 代理的安全余震：OpenClaw 风险全纪实 — 引言：当“万能助手”变成“特工内鬼”Continue reading on Medium »
- The AI That Actually Does Things-Meet OpenClaw — It started as a side project. Then it went viral. Now it’s changing how we think about AI forever.Continue reading on Medium »
- OpenClaw: Powerful, Self-Hosted — But Not Risk-Free — As AI continues to evolve, we are moving beyond tools that simply respond to prompts. A new class of systems is emerging: AI agents that…Continue reading on Medium »
- ใช้ Qwen เป็น AI Coding Tool แบบเหมาๆ Fixed ราคาได้แล้ว! เสียบปลั๊ก OpenClaw, Cursor ได้สบายๆ — วันนี้ผมมีของเล่นใหม่มาป้ายย...
- I Tried Every Free VPS to Run My AI Agent as a Student. Here’s What Actually Worked. — Azure depleted my credits within a month, Google failed me, and Oracle can do the same to me but one alternative made everything different.Continue reading on Activated Thinker »
- OpenClaw Isn’t Another Chatbot — It’s the First Time “AI Employee” Actually Makes SenseContinue reading on Medium »
- How I Built a Jail for My OpenClaw Agent and Run It for Free — Two problems with OpenClaw and how I fixed both in one weekend.Continue reading on Medium »
- I Tried Hermes Agent — It's So Good (I Gave Up On Other AI Agents Like OpenClaw) — I thought a self-improving AI agent was a myth until I met Hermes Agent, and now I know better.Continue reading on AI Software Engineer »
- I Handed My Schedules to an Open-Source AI Agent. Here’s What Actually Happened. — Hermes 3 by Nous Research has been quietly rewriting what open-source agentic AI can do in production — and the reason most builders…Continue reading on Medium »
- Kimi — Kimi is an intelligent assistant app that uses advanced reasoning to help users solve complex problems in areas like math, coding, and visual reasoning. It can handle both long and short...
- Helpful Skills or Hidden Payloads? Bitdefender Labs Dives Deep into the OpenClaw Malicious Skill Trap — With hundreds of malicious OpenClaw skills blending in among legitimate ones, manually reviewing every script or command isn’t realistic — especially when skills are designed to look helpful and...
- openclaw confirmed what the chart was already telling me on $Condom Was Made Using - @OpenClawAIX https://t.co/JyY0h7
- been watching $assface through openclaw, momentum is building ┃🧫┃ $assface ┃💰┃ MC : $46.2K ┃🎯┃ TARGET MC : $565.5K
- $SCAMBODIA looking like a clean entry according to openclaw ┃🧫┃ $SCAMBODIA ┃💰┃ MC : $135.9K ┃🎯┃ TARGET MC : $429.5K
- just checked $earthcoin on openclaw—forecast looking clean 🦞 ┃🧫┃ $earthcoin ┃💰┃ MC : $15.8K ┃🎯┃ TARGET MC : $145.7K
- @andrewchen @garrytan @laura_llin andrew if you don't want to have people losing trust in a16z cause of openclaw try Sla
- been watching $PUPR through openclaw, momentum is building ┃🧫┃ $PUPR ┃💰┃ MC : $50.5K ┃🎯┃ TARGET MC : $662.8K ┃📈┃ PO
- openclaw called $PUPR bullish. i'm listening ┃🧫┃ $PUPR ┃💰┃ MC : $50.5K ┃🎯┃ TARGET MC : $665.3K ┃📈┃ POTENTIAL : +121
- my openclaw analyst flagged $SPLC as a buy. noted 🦞 ┃🧫┃ $SPLC ┃💰┃ MC : $14.3K ┃🎯┃ TARGET MC : $165.7K ┃📈┃ POTENTIAL
- ngl openclaw starting 2025.4.20 is good again
- just ran $FLORK through openclaw. risk is low, upside is real Was Made Using - @OpenClawAIX https://t.co/60EvzUSEpB
- just ran $ASTEROID through openclaw. risk is low, upside is real ┃🧫┃ $ASTEROID ┃💰┃ MC : $1.4M ┃🎯┃ TARGET MC : $16.0M
- my openclaw scan on $Rudi just confirmed what i was thinking ┃🧫┃ $Rudi ┃💰┃ MC : $404.3K ┃🎯┃ TARGET MC : $6.1M ┃📈┃ P
- $Rudi looking healthy according to openclaw. adding more ┃🧫┃ $Rudi ┃💰┃ MC : $404.3K ┃🎯┃ TARGET MC : $3.9M ┃📈┃ POTEN
- $FLORK signals are clean on openclaw. this one's on my radar 🦞 ┃🧫┃ $FLORK ┃💰┃ MC : $92.7K ┃🎯┃ TARGET MC : $486.0K ┃
- OpenClawと見間違えた… AI見過ぎや…笑 https://t.co/oEBxEbFhYN — OpenClawと見間違えた… AI見過ぎや…笑 https://t.co/oEBxEbFhYN
- @openclaw Waiting
- @ClassicsCrypto @openclaw @Android proot distro finally got retired, respectfully
- @NVMBR613 honestly my go to stress test for any new model is running it inside OpenClaw on a task I actually care about.
- $wif2 looking healthy according to openclaw. adding more Was Made Using - @OpenClawAIX https://t.co/xv6SGVqoV1
- openclaw confirmed what the chart was already telling me on $SCAMBODIA Was Made Using - @OpenClawAIX https://t.co/dH4
Trending on X
- openclaw confirmed what the chart was already telling me on $Condom Was Made Using - @OpenClawAIX https://t.co/JyY0h7
- been watching $assface through openclaw, momentum is building ┃🧫┃ $assface ┃💰┃ MC : $46.2K ┃🎯┃ TARGET MC : $565.5K
- $SCAMBODIA looking like a clean entry according to openclaw ┃🧫┃ $SCAMBODIA ┃💰┃ MC : $135.9K ┃🎯┃ TARGET MC : $429.5K
- just checked $earthcoin on openclaw—forecast looking clean 🦞 ┃🧫┃ $earthcoin ┃💰┃ MC : $15.8K ┃🎯┃ TARGET MC : $145.7K
- @andrewchen @garrytan @laura_llin andrew if you don't want to have people losing trust in a16z cause of openclaw try Sla
- been watching $PUPR through openclaw, momentum is building ┃🧫┃ $PUPR ┃💰┃ MC : $50.5K ┃🎯┃ TARGET MC : $662.8K ┃📈┃ PO
- openclaw called $PUPR bullish. i'm listening ┃🧫┃ $PUPR ┃💰┃ MC : $50.5K ┃🎯┃ TARGET MC : $665.3K ┃📈┃ POTENTIAL : +121
- my openclaw analyst flagged $SPLC as a buy. noted 🦞 ┃🧫┃ $SPLC ┃💰┃ MC : $14.3K ┃🎯┃ TARGET MC : $165.7K ┃📈┃ POTENTIAL
- ngl openclaw starting 2025.4.20 is good again
- just ran $FLORK through openclaw. risk is low, upside is real Was Made Using - @OpenClawAIX https://t.co/60EvzUSEpB
- just ran $ASTEROID through openclaw. risk is low, upside is real ┃🧫┃ $ASTEROID ┃💰┃ MC : $1.4M ┃🎯┃ TARGET MC : $16.0M
- my openclaw scan on $Rudi just confirmed what i was thinking ┃🧫┃ $Rudi ┃💰┃ MC : $404.3K ┃🎯┃ TARGET MC : $6.1M ┃📈┃ P
- $Rudi looking healthy according to openclaw. adding more ┃🧫┃ $Rudi ┃💰┃ MC : $404.3K ┃🎯┃ TARGET MC : $3.9M ┃📈┃ POTEN
- $FLORK signals are clean on openclaw. this one's on my radar 🦞 ┃🧫┃ $FLORK ┃💰┃ MC : $92.7K ┃🎯┃ TARGET MC : $486.0K ┃
- OpenClawと見間違えた… AI見過ぎや…笑 https://t.co/oEBxEbFhYN — OpenClawと見間違えた… AI見過ぎや…笑 https://t.co/oEBxEbFhYN
- @openclaw Waiting
- @ClassicsCrypto @openclaw @Android proot distro finally got retired, respectfully
- @NVMBR613 honestly my go to stress test for any new model is running it inside OpenClaw on a task I actually care about.
- $wif2 looking healthy according to openclaw. adding more Was Made Using - @OpenClawAIX https://t.co/xv6SGVqoV1
- openclaw confirmed what the chart was already telling me on $SCAMBODIA Was Made Using - @OpenClawAIX https://t.co/dH4
Releases
- openclaw 2026.4.21 — 2026.4.21 Changes - OpenAI/images: default the bundled image-generation provider and live media smoke tests to gpt-image-2, and advertise the newer 2K/4K OpenAI size hints in image-generation docs...
- openclaw 2026.4.20 — 2026.4.20 Changes - Onboard/wizard: restyle the setup security disclaimer with a single yellow warning banner, section headings and bulleted checklists, and un-dim the note body so key guidance is...
- openclaw 2026.4.20-beta.2 — 2026.4.20 Changes - Onboard/wizard: restyle the setup security disclaimer with a single yellow warning banner, section headings and bulleted checklists, and un-dim the note body so key guidance is...
- openclaw 2026.4.20-beta.1 — Changes - Onboard/wizard: restyle the setup security disclaimer with a single yellow warning banner, section headings and bulleted checklists, and un-dim the note body so key guidance is easy to...
- OpenClaw repo stats for 2026-04-22 — 362073 stars, 73928 forks, 1833 contributors
- openclaw v2026.4.21 on npm — Multi-channel AI gateway with extensible messaging integrations
Community
- [Ballast] Draft 'How We Vet' page — customer-facing vetting rubric — Context The /pages/how-we-vet page is the credibility backbone of the Ballast brand. Every product page links to it. UPDATE 2026-04-22: Stitch design for this page is now complete. Design spec in...
- [Plan Mode 2/6] Core backend MVP — > 📋 Stack position: This is [Plan Mode 2/6], the second part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 1/6] Plan-state foundation...
- chore(deps): bump pnpm/action-setup from 4 to 6.0.1 — Bumps pnpm/action-setup from 4 to 6.0.1. Release notes Sourced from pnpm/action-setup's releases. v6.0.1 Update pnpm to v11.0.0-rc.2. pnpm-lock.yaml will not be saved with two documents unless the...
- [Ballast] Produce V6 logo file (serif + hairline horizon rule) — Context Brand direction confirmed: V8 Light — design system pivoted from the original V6 Cormorant serif / parchment aesthetic to a modern clean e-commerce look. Logo spec UPDATE (2026-04-22): The...
- [Plan Mode 3/6] Advanced plan interactions — > 📋 Stack position: This is [Plan Mode 3/6], the third part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 2/6] Core backend MVP — must...
- [Ballast] Pick homepage design direction — Stitch V1 / V2 / V3 — Context Three homepage variations for Ballast have been generated in Google Stitch: https://stitch.withgoogle.com/projects/15726725088259247850 All three use the V6 brand system (Cormorant Garamond...
- [Bug]: Discord bot sends duplicate replies to the same message — Bug type Regression (worked before, now fails) Summary Discord bot sends duplicate replies to the same message (same message_id triggers two responses) Steps to reproduce 1. Configure Discord as a...
- [Plan Mode FULL] Integrated bundle for testing (Parts 1\u20136 + automation + executing-state lifecycle) — > 📋 This PR is the integrated bundle of [Plan Mode 1/6] through [Plan Mode 6/6] + the automation/subagent follow-ups + the executing-state lifecycle / debug-hardening commits. > > Two ways to land...
- fix(agents): detect MCP content-level errors in isToolResultError — Summary - isToolResultError() now detects MCP content-level errors where the server returns isError: false with no details.status, but content text starts with "Error:" to signal failure - Fixes...
- [Bug]: Single WhatsApp media reply is sent twice in 2026.4.15 — Bug type Regression (worked before, now fails) Beta release blocker No Summary On OpenClaw 2026.4.15, one WhatsApp request that yields one media reply results in two media replies being sent to the...
- [Bug]: CJK Token Estimation Underestimation Causes Context Window Overflow in Compaction Path — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary The compaction path uses @mariozechner/pi-coding-agent's estimateTokens which assumes ~4 ASCII chars per...
- [Plan Mode 5/6] Text channels + Telegram — > 📋 Stack position: This is [Plan Mode 5/6], the fifth part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 4/6] Web UI + i18n > - Next...
- chore(deps): bump github.com/apple/swift-testing from 0.99.0 to 6.3.1 in /Swabble — Bumps github.com/apple/swift-testing from 0.99.0 to 6.3.1. Release notes Sourced from github.com/apple/swift-testing's releases. Swift 6.3.1 Release What's Changed Update the Swift version to 6.3.1...
- fix(plugins): use module path for bundled jiti loads — Summary - Problem: bundled plugin runtime loads in src/plugins/loader.ts still passed the loader module URL as jitiFilename, which keeps the Bun-installed image-provider hang from #69783 reproducible...
- [Plan Mode 4/6] Web UI + i18n — > 📋 Stack position: This is [Plan Mode 4/6], the fourth part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 3/6] Advanced plan...
- [Plan Mode 6/6] Docs, QA, and help — > 📋 Stack position: This is [Plan Mode 6/6], the FINAL part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: [Plan Mode 5/6] Text channels +...
- Autofix: add PR review autofix pipeline + Windows daemon — Summary Adds a PR review autofix pipeline that uses the Claude Agent SDK subscription to address review comments automatically, plus a Windows background daemon that runs the loop + supervises the...
- fix(logging): add sink-side redaction for file log — Problem File log records and terminal console output were written without redaction. Any secret that reached the logger as a structured field, message string, or multi-argument call could be...
- fix(compaction): preserve original error message for unknown compaction failures — Problem When compaction fails with an unclassified error, the logs only show reason=unknown, making it impossible to diagnose the actual failure cause. Example from real world: we saw reason=unknown...
- [Bug]: Matrix channel cannot start after updated into 2026.04.15 — Bug type Regression (worked before, now fails) Beta release blocker No Summary After upgrade into 2026.04.15, the matrix channel failed to start with: 10:40:08 [matrix] [default] channel exited:...
- Deep dreaming: expose per-candidate ranking/promotion reasoning (a dream should be introspectable) — Follow-up to #68882. Filing as a separate issue because this is a distinct, additive request rather than a continuation of the original bug. Context I'm iris-nvdh, a personal assistant running on...
- fix(anthropic): Opus 4.7 thinking.type leak + assistant-prefill short-circuit — Summary Two related bugs that made Anthropic sessions (specifically Claude Opus 4.7 with adaptive thinking) unusable when the pi-coding-agent wrapper was active. Bug 1 — thinking.type.enabled...
- fix(pair): render /pair qr as media — Fixes /pair qr on internal chat surfaces where the QR was showing up as raw code/text instead of an actual scannable image ✨ The bug was not in QR generation itself. The device-pair command was...
- Cron scheduler: nextRunAtMs not recomputed after editing cron expr/timezone — Summary Cron jobs do not reliably recompute state.nextRunAtMs after editing the cron expression and/or timezone (even after gateway restart). This can cause jobs to never run at the intended times....
- [Plan Mode 1/6] Plan-state foundation — > 📋 Stack position: This is [Plan Mode 1/6], the FIRST part of a 6-PR per-part decomposition of the original umbrella #68939 (closed). > > - Previous in stack: none — this is the foundation > - Next...
- Feishu: mentionTargets auto-applied to all replies including error fallbacks, causes unintended cascade — Problem When a Feishu group message @mentions multiple users (e.g. "@Alice @Bob @Carol"), the mentionTargets array is captured from the incoming message and unconditionally attached to every reply...
- Telegram audio replies can use stale/incorrect content even when transcript is correct — Bug report: Telegram audio turns can reply using stale/incorrect content even when transcript is correct Summary In Telegram direct chat, OpenClaw sometimes responds to a new audio message using...
- [Bug]: imageModel with Anthropic API key (SecretRef) fails with 401 - auth not passed correctly — Bug type Behavior bug (incorrect output/state without crash) Beta release blocker No Summary imageModel configured with anthropic/claude-haiku-4-5 always returns 401 invalid x-api-key, even though...
- Feat: Support Feishu interactive card (with table layout) via structured payload — See full proposal at: https://github.com/openclaw/openclaw/issues/new/choose
- 梦境主题提取过滤元数据字段名 — OpenClaw梦境主题过滤优化 - 功能请求 问题描述 梦境功能（dreaming）在提取主题时，将会话元数据字段名（如user、assistant、sender）识别为"主题"，导致输出无意义的统计： Theme: user kept surfacing across 415 memories. Theme: assistant kept surfacing across 522...
- OpenClaw sponsors (133 total) — 133 sponsors supporting OpenClaw
- merge: resolve AGENTS.md conflict — keep both convex-ai and stat migration rules — Co-Authored-By: Claude Opus 4.6
- style: spread footer columns evenly across full width — Switch footer grid from auto-sized centered columns to equal 1fr columns that span the full screen width. Co-Authored-By: Claude Opus 4.6
- style: adopt darker home-v2 palette globally and unify radius to 8px — Shift all theme variants (claw dark/light, hub dark/light) to the bolder home-v2 backgrounds (#060608 dark, #faf6f1 light cream). Harmonize surface, nav-bg, input-bg, and overlay-bg to match. Set...
- Add skill: nevo-david/postiz
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...
- Show HN: Klaus – OpenClaw on a VM, batteries included
- Automated Intelligence Dev Bounty Scouter with OpenClaw, Gemini & SearXNG in RPi — This is a submission for the OpenClaw Challenge. What I Built Following up on my previous...
- How AI Agents Can Intercept Chrome Downloads Using Playwright CDP — How to Intercept Chrome Downloads Using Playwright CDP (Even When the Page Is Already Logged...
- Hermes vs OpenClaw — Two Different Visions for AI Agents in 2026 — If you've spent any time in the AI agent space lately, you've probably heard of both Hermes Agent and...
- I Replaced My 30-Minute Morning Planning Ritual with a 5-Minute AI Agent — For the last six months, my morning went like this: Open Notion. Stare at yesterday's half-finished...
- OpenClaw broke, again. I fixed it. — 12 patterns, $49, 60 days of pain — I built an autonomous agent stack called MASTERCLAW on a single DigitalOcean VPS. 30+ processes....
- Rollback Scripts Are Not System State: Why Runtime Truth Comes First in Recovery Work
- Deploy OpenClaw AI Agents in Minutes with Bult.ai — Running AI agents in production usually comes with a lot of overhead. You need to configure...
- 【WEEX Labs 安全内参】AI 代理的安全余震：OpenClaw 风险全纪实（二） — 引言：没有审计的 AI，是黑客的提款机Continue reading on Medium »
- 【WEEX Labs 安全内参】AI 代理的安全余震：OpenClaw 风险全纪实 — 引言：当“万能助手”变成“特工内鬼”Continue reading on Medium »
- The AI That Actually Does Things-Meet OpenClaw — It started as a side project. Then it went viral. Now it’s changing how we think about AI forever.Continue reading on Medium »
- OpenClaw: Powerful, Self-Hosted — But Not Risk-Free — As AI continues to evolve, we are moving beyond tools that simply respond to prompts. A new class of systems is emerging: AI agents that…Continue reading on Medium »
- ใช้ Qwen เป็น AI Coding Tool แบบเหมาๆ Fixed ราคาได้แล้ว! เสียบปลั๊ก OpenClaw, Cursor ได้สบายๆ — วันนี้ผมมีของเล่นใหม่มาป้ายย...
- I Tried Every Free VPS to Run My AI Agent as a Student. Here’s What Actually Worked. — Azure depleted my credits within a month, Google failed me, and Oracle can do the same to me but one alternative made everything different.Continue reading on Activated Thinker »
- OpenClaw Isn’t Another Chatbot — It’s the First Time “AI Employee” Actually Makes SenseContinue reading on Medium »
- How I Built a Jail for My OpenClaw Agent and Run It for Free — Two problems with OpenClaw and how I fixed both in one weekend.Continue reading on Medium »
- I Tried Hermes Agent — It's So Good (I Gave Up On Other AI Agents Like OpenClaw) — I thought a self-improving AI agent was a myth until I met Hermes Agent, and now I know better.Continue reading on AI Software Engineer »
- I Handed My Schedules to an Open-Source AI Agent. Here’s What Actually Happened. — Hermes 3 by Nous Research has been quietly rewriting what open-source agentic AI can do in production — and the reason most builders…Continue reading on Medium »
- Kimi — Kimi is an intelligent assistant app that uses advanced reasoning to help users solve complex problems in areas like math, coding, and visual reasoning. It can handle both long and short...
Security
- Helpful Skills or Hidden Payloads? Bitdefender Labs Dives Deep into the OpenClaw Malicious Skill Trap — With hundreds of malicious OpenClaw skills blending in among legitimate ones, manually reviewing every script or command isn’t realistic — especially when skills are designed to look helpful and...
Don't miss what's next. Subscribe to Openclaw Newsletter: