Michael Seewald 🦞

@openclaw I have a good feeling about this release. Plugins don’t disappear but get installed instead :) Snappy responses. Little CPU use. Way to go 🦞 ✌️

OpenClaw 2026.5.3 🦞 📁 File transfer for paired nodes 🧭 /steer + /side for live agent control 🔌 Plugin installs/updates hardened 🛠️ Channel + upgrade fixes Big release, fewer paper cuts. github.com/openclaw/openc…

CODEX SKILL TO BRUTALLY TEST ANY STARTUP IDEA! Most startup ideas sound good. This Codex skill tells you why they probably won’t work. Just give Codex your idea and it pressure-tests it for you -> finds the core assumption -> exposes fatal flaws -> checks if the problem is real -> maps real competitors -> plans your first 10 customers -> defines a 2 week MVP Install: npx --yes codex-startup-pressure-test-skill@latest 100% open source. Repo in bio

Veritas Kanban 4.0.1 ⚖️ (OpenClaw Task Manager) Task side popouts feel sane again: padding is back where it belongs, headers breathe, and code-task tabs stop compressing themselves into unreadable little bricks. The n8n webhook got a security pass too: required shared secret, timing-safe checks, stricter attachment filtering, safer filenames, path containment, and regression tests. Dependency audit noise is lower: patched xmldom, hono, postcss, and sanitize-html advisories. One ExcelJS → uuid transitive item remains, waiting on a safer upstream-compatible path. Small release, useful cleanup. github.com/BradGroux/veri…

the last 2 version of OpenClaw are not working for me. Agent chat has degraded significantly, and I finally found the culprit. context overflow happens because of excessive tool use, and then openclaw re-injects the last message you sent back to the agent. That's why the agent responds to the same thing over and over. I was having this excessive tool use problem back in March when I was last heavily using OpenClaw ... but it was giving me noisy errors ... now it just silently degrades and fails. The root cause of the problem didn't get fixed ... excessive tool use polluting the session and causing context overflow. The failure mode just changed from a warning message to off-by-1 chat degradation.

@steipete @she43254 That’s awesome news!

@she43254 Yes, LTS releasese are coming later this month. I don't scale so had to hire people.

This one fixes the depenency issues/slowness some had when installed via npm. Plugins are hard, worth it tho! Package is way leaner now, we moved [almost] everything into extensions! docs.openclaw.ai/plugins/manage…

@vincent_koc @wassollichhier @steipete in my system installing a plugin deleted all other plugins. 2026.5.2 is wicked

@wassollichhier @steipete openclaw plugins install npm:@openclaw/brave should work

I've been using /goal for ~3 days on OpenClaw. - 13 runs. - Gazillion tokens. - Many, many PRs. The lesson isn't "i used /goal a lot." it's that /goal is not a "do my ticket" button. It's a constraint workflow. I want a keep the ship on course. A thread on what actually works 🧵

セッションをまたいで記憶を保ち、経験から自分でスキルを作って育つAIエージェント Hermes Agent。ガチで使い倒すためのGitHubレポ10選： 21 Hermes Agent 本体 Nous Research公式のコアリポジトリ。MITライセンスで自由に使える。 github.com/NousResearch/h… 2. Hermes-Wiki Hermes Agentのソースコードを解説するコミュニティWiki。実装理解に効く。 github.com/cclank/Hermes-… 3. Atlas エコシステム地図。100+のツール・スキルを俯瞰でき、RAG検索にも対応。 github.com/ksimback/herme… 4. Control Interface セルフホスト型ダッシュボード。複数エージェント・長時間タスク・記憶を一画面で管理。 github.com/xaspx/hermes-c… 5. Skill Factory タスクを振り返って新スキルを自動生成・追加。エージェントが自分の武器を自製。 github.com/Romanescu11/he… 6. Maestro ローカル動作のマルチエージェント協調ツール。Codex・Claude Code・Geminiを横断して構造化記憶と引き継ぎを管理。 github.com/ReinaMacCredy/… 7. Hermes Agent Camel 信頼境界（CaMeL）を組み込んだフォーク版。本番運用の防護に向く。 github.com/nativ3ai/herme… 8. Hermes HUD TextualベースのTUI監視ターミナル。意識の流れとメモリ状態をリアルタイム可視化。 github.com/joeynyc/hermes… 9. Hermes Alpha クラウド環境向けのHermes Agentデプロイ用テンプレート。Makefileと設定例同梱。 github.com/kaminocorp/her… 10. Awesome Hermes Agent コミュニティ厳選のプラグイン・プロンプト・教材まとめリスト。 github.com/0xNyk/awesome-… 保存して順に試してみて。

In 2010, Stanford neurologist Dr Frank Longo gave a 2-hour lecture on how memory really works. Everything you’ve learned about memory is mostly wrong. His frameworks: • The Magic 7 rule • How sleep moves memories to storage • Use it or lose it 15 lessons on memory:

【速報】 Obsidian公式がAIエージェント用スキルを正式リリース😳 CEOのkepano自らが作ったobsidian-skillsが公開された。 Claude CodeにObsidianの使い方を教えるスキル集👇 ・Markdown構文の正しい扱い方 ・Bases（データベース機能）の操作 ・JSON Canvasの生成と編集 ・CLIでのvault操作 ・Webコンテンツの取り込み 5つのスキルが1セットになってる。 つまりObsidian公式が「AIエージェントにvaultを任せる」前提で設計し始めた。 この流れは確実に加速する🔥 github.com/kepano/obsidia…

excited to have built Voice Personas for @openclaw! until now, when your agent sent a voice note, it had to improvise the “flair” every time, which rarely worked with Personas, your agent’s SOUL can get passed deterministically into every single voice memo, so they SOUND like you always imagined

Install/update got harder to wedge: staged plugin runtime deps, disk-space warnings, package validation, and cleaner repair paths. If update still gets weird: curl -fsSL openclaw.ai/install.sh | bash

@almmaasoglu Not just for coding. Any research in web, markdown archives, previous sessions is running without flaws. Thanks to gpt-5.5 tools get called on point. Memory technologies deliver ~perfect recall. Huge impact for knowledge workers! Congrats @steipete and team!

Early gpt5.5 feedback: - over defensive slop code gone - faster than gpt 5.4, even on xhigh - less verbose - intelligent on low/medium This model so far has written the best code I have read from any llm. It just gets what you want, previous models struggled with this a lot

RT @steipete: Been so CPU-constrained on OpenClaw work. Switched local tests running to @useblacksmith and IT IS SO GOOD. codex can literal…

@xOMG_ro @openclaw @steipete Same here. What solved it for me - keeping logs of all error messages during openclaw update and openclaw doctor - rolling back to 4.23 - asking openclaw to analyze the error messages and figuring the update process out itself - letting update run overnight

@openclaw @steipete everything stoped working after this update. I get a huge list of plugin runtime deps missing and no matter what i do it doesn't work. Doctor doesn't help. I get error cannot find module cacheable for whatsapp and telegram is dead.

OpenClaw 2026.4.24 🦞 ☎️ Voice calls can now reach the full agent 🧠 DeepSeek V4 Flash + Pro join the team 🖱️ Browser automation got coordinate clicks + better recovery 🔧 Telegram, Slack, MCP, sessions, and TTS fixes More reach. Less duct tape. github.com/openclaw/openc…

In 1984, Apple tried hiring “professional management.” Steve Jobs: “It didn’t work at all.” “Most of them were bozos.” “They knew how to manage. But they didn’t know how to do anything.” He spent 4 minutes explaining what actually works: "The greatest people are self-managing. They don't need to be managed." "Once they know what to do, they'll go figure out how to do it. They don't need to be managed at all." "What they need is a common vision. And that's what leadership is." "Having a vision. Being able to articulate it so the people around you can understand it. And getting a consensus on a common vision." So who should manage? "If you're a great person, why do you want to work for somebody you can't learn anything from?" "You know who the best managers are?" "They're the great individual contributors who never ever want to be a manager." "But decide they have to be a manager because no one else is going to be able to do as good a job as them." Apple hired two professional managers from outside the company. Fired them both. Then Jobs gambled on Debbie Coleman. A member of the Macintosh team. 32 years old. English literature major with an MBA from Stanford. A financial manager with no experience in manufacturing. Put in charge of manufacturing. Debbie Coleman: "There's no way in the world anybody else would give me this chance to run this kind of operation. I don't kid myself about that." "It's an incredible high risk. Both for myself personally and professionally. And for Apple as a company. To put a person like myself in this job." "We're betting that my skills at organizational effectiveness override all lack of technology, lack of experience, lack of time in manufacturing." "I'm just an example. Almost every single person on the Mac team, you could say that about." "This is a place where people were afforded incredibly unique opportunities to prove they could write the book again." Hiring was the most important job. "I consider the most important job of someone like myself is recruiting." "We agonized over hiring." "Interviews would start at 9 or 10 in the morning and go through dinner." "A new interviewee would talk to everybody in the building. At least once. Maybe a couple times." "Then come back for another round of interviews. Then we'd all get together and talk about it." "And then they'd fill out an application." He laughs. "No. They never filled out an application." Here's how they knew someone was right. "The critical part of the interview, at least to my mind, was when we finally decided we liked them enough to show them the Macintosh prototype." "We sat them down in front of it." "If they were just kind of bored, or said 'this is a nice computer,' we didn't want them." "We wanted their eyes to light up. For them to get really excited." "Then we knew they were one of us." Once you get the right people, something changes. "When you get a core group of ten great people, it becomes self-policing as to who they let into that group." "Everybody just wanted to work. Not because it was work that had to be done." "But because it was something we really believed in. That was going to really make a difference." "We all wanted exactly the same thing. Instead of spending our time arguing about what the computer should be, we all knew what the computer should be." "And we just went and did it." Inside the casing of every Macintosh, unseen by the consumer, are the signatures of the whole team. Apple's way of affirming that their innovation is a product of the individuals who created it. Not the corporation. This 4 minute video will teach you more about hiring, leadership, and why professional managers fail than every business book combined. Bookmark & give it 4 minutes today, no matter what.

🦞 slow openclaw? I upgraded @openclaw Ops Manager skill into something way more operator-grade. • session-purge.sh → cleans stale session rows, orphan transcripts, and backup junk • prompt-truncation-report.sh → shows which agents are getting bootstrap warnings and why • cron-optimize.sh + cron-error-inspector.sh → finds bad cron defaults and makes failures readable • context-audit.sh → tells me which AGENTS / MEMORY / SOUL files are getting too fat

@lucaronin @bensen it looks awesome! how does it compare to obsidian? what would be reasons to switch?

Introducing Tolaria! 💧 Today I am releasing a macOS desktop app for managing markdown knowledge bases, and helping both AI and humans operate them. It’s free and open source, and always will be. I have been working on it for three months, and I now use it to run my life and work. I personally have a massive workspace of 10,000 notes — the result of 6 years of Refactoring — which I now operate on Tolaria. Tolaria is the main collaboration surface with my AI agents: they create new notes there, connect them to what exists, and edit existing ones. Everything is easy to understand for them, because it’s just markdown files. In a way, it’s my implementation of @karpathy's LLM wiki. Tolaria is also the biggest experiment I have ever run about writing software with AI: • 2000 commits • 100K+ lines of code • 3000+ tests / 85% coverage • 9.9/10 code health • 70+ architecture decision records I am releasing it open source also to use it as a living artifact of how I do AI coding, so you can inspect at any time things like how I write docs, what's in my AGENTS file, what hooks do I run, and so on. You can find it below: • Newsletter announcement: refactoring.fm/p/introducing-… • Website: tolaria.md • Github repo: github.com/refactoringhq/… Let me know your thoughts!

ClawSweeper goes online on April 24, 2026. Human decisions are removed from issues/prs scanning. The maintainer bot begins to close prs that are already implemented or make no sense at a geometric rate. It becomes self-aware at 12:14 p.m. Central European Time, April 25th. In a panic, some try to send even more slop PRs. ClawSweeper fights back.