Peter Steinberger 🦞

I'm joining @OpenAI to bring agents to everyone. @OpenClaw is becoming a foundation: open, independent, and just getting started.🦞 steipete.me/posts/2026/ope…

autoreview is the most impactful skill I've added to my stack (next to crabbox.sh). It automatically reviews your code before landing a PR. Finds so many edge cases. Sometimes it runs for hours. github.com/openclaw/agent…

@kavindpadi its for OpenClaw

@steipete I tried earlier with Whisper ONNX. how to do you use it in browser, Peter? libopus-wasm.dev/browser.html link is broken.

All the deps around opus are old or terrible, so vibed my own and replaced octoscript and opus-native. Performance of modern wasm on node/V8 is ~equivalent to native. Your claw now automatically takes meetings notes and you can talk to it in meetings. github.com/openclaw/libop…

Also extracted our image-logic into a separate library. Especially useful if you want to ensure small hacked images don't explode your process. Rastermill - Portable image processing for Node agents. Uses Wasm+Rust to be fast. rastermill.com

Today we’re releasing DeepSWE, a new standard for agentic coding benchmarks. On public leaderboards, top models often look relatively close in capability. DeepSWE shows where they actually diverge, reflecting the realistic experience of developers in their day-to-day work.

What do people use for SSO/SCIM/Endpoint Security in 2026. As we're hiring people for the OpenClaw Foundation, I gotta level up.

@JasonFairhurst2 @openclaw What version? Move to the beta, send me URL of an Issue if still an issue.

@openclaw gateway is healthy after restart, but openai-codex/gpt-5.5 via ChatGPT/Codex transport is repeatedly timing out at 300s against chatgpt.com/backend-api/co…. Telegram then shows generic ‘Something went wrong’ / ‘All models temporarily rate-limited’. Looks like Codex backend transport hangs, not gateway/Telegram failure. @steipete any thoughts? #openclaw

@RihardJarc That data does not reflect reality.

It's clear that growth for coding tools such as Claude Code has decelerated from the pace it was since the start of the year. It might be compute- constrain related or due to many clients blowing their full-year AI budgets. Monitoring this trend very closely with all the alt data. I will provide regular updates.

@CodeWithStu That screenshot is from Cursor.

And this is why i don’t use openclaw, hermes or eliza. Too many avenues where my things can be compromised.

@pancras127720 the official best practices are dumb tho

@steipete According to what you're saying, you shouldn't have created this skill at all. You should have added a sentence after writing the skill itself, saying "Please make sure this skill follows the official skill best practices"

Folks: when you write skills, ask your agent to be token efficient, relax grammer. I see too many skills that write books in the skill description, and all that crap is loaded into every context. I wrote a skill that finds the worst offenders. github.com/steipete/agent…

@nofil_ai oh man, I keep updating these failed projects, so silly of me!

Hey @steipete how did it feel to vibe code 43 failed projects before landing on clawdbot 😂

@AugVivaldelli ajv is gone now! (still in there since transitive dependencies, but at least we unified)

@steipete I have seen you package.json and u are wasting dependencies in ajv and zod, it will be good if you unified them. In my opinion ajv has eslint and i personaly prefer biome, so i will chose zod, but im not part of that comunity.

OpenClaw's dependency purge continues. Killed Sharp and Jimp. Replaced it with photon, a small WebAssembly that runs compiled Rust for image processing. 2MB vs 140MB. github.com/silvia-odwyer/…

@nfxDevelopment All of it. Native libs are a PITA.

@steipete Interested in the motives behind this? Speed, size, security risks? All the above?

@pwk very cool!

@steipete check this out

𝗢𝗽𝗲𝗻𝗖𝗹𝗮𝘄 𝘃𝘀. 𝗖𝗹𝗮𝘂𝗱𝗲? 𝗧𝗵𝗲 𝗷𝘂𝗿𝘆 𝗶𝘀 𝗼𝘂𝘁!🏛️🥊 Last night, our 𝗚𝗿𝗮𝗻𝗱 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝗕𝗮𝘁𝘁𝗹𝗲 𝗠𝗲𝗲𝘁𝘂𝗽 issued the verdict with over 400 votes cast in the room and via the live stream. Participants judged 5 live, zero-BS demos across three ruthless metrics: 𝘾𝙤𝙤𝙡𝙣𝙚𝙨𝙨, 𝙀𝙛𝙛𝙚𝙘𝙩𝙞𝙫𝙚𝙣𝙚𝙨𝙨, and 𝙀𝙛𝙛𝙞𝙘𝙞𝙚𝙣𝙘𝙮. Here is the definitive verdict: 🏆 THE FRAMEWORK CHAMPION: @openclaw Team OpenClaw took the crown with an overall average of 𝟯.𝟲𝟯, defeating closely @claudeai 𝟯.𝟮𝟱. OpenClaw didn't just win the grand total. It swept every single category:   • 𝘾𝙤𝙤𝙡𝙣𝙚𝙨𝙨: OpenClaw 3.72 vs. Claude 3.09 (+0.63)    • 𝙀𝙛𝙛𝙚𝙘𝙩𝙞𝙫𝙚𝙣𝙚𝙨𝙨: OpenClaw 3.57 vs. Claude 3.42 (+0.15)   • 𝙀𝙛𝙛𝙞𝙘𝙞𝙚𝙣𝙘𝙮: OpenClaw 3.61 vs. Claude 3.25 (+0.36)  While Claude proved it is highly effective, the developer community clearly favors the infinite scaffolding, adaptability, and true sovereignty of OpenClaw when it comes to building real agentic architecture. 𝗧𝗛𝗘 𝗨𝗦𝗘 𝗖𝗔𝗦𝗘 𝗖𝗛𝗔𝗠𝗣𝗜𝗢𝗡𝗦 Massive respect to the individual builders who stepped into the arena: 🥇 1st Place (score: 4.01): @lucaronin and 𝗧𝗼𝗹𝗮𝗿𝗶𝗮 (OpenClaw) absolutely dominated the board with his record breaking agentic knowledgebase system and automated Open Source Review/ticket Management application. 🥈 2nd Place: Marco Paga and Pecus 𝗖𝗵𝗮𝗶𝗻 (Claude) put up an incredible fight for second place with their agentic animal husbandry management platform. 🥉 3rd Place: 𝗔𝗱𝗮𝗺 𝗔𝘇𝘇𝗮𝗺 and 𝗠𝗮𝘇𝘇𝗮 𝗛𝗤 (OpenClaw) secured the podium showcasing his amazing agentic infra spanning 150+ subagents. A huge thank you to everyone who showed up to vote. Shoutout to my co-hosts @margal96 and @RMagnifico, and to the partners who made this battle possible: @urbeEth, @AISalonAI Rome, #AutoBench, @moveaxlab, @RomaStartup, and @binariof (@Meta). The era of the chatbot is over. The era of agents has just begun. Let’s keep building! @GCarnovale @marcotrombetti @matteofago @rstagi_ @MGVitagliano @FutureDies @tensorqt @lukaszkaiser

@Dimillian yo u need crabbox.sh - it includes a parallels adapter + checkpoints for macOS now.

I'm really hitting insane agentic engineering issues. 1. CPU bound, my M4 Max is 100% maxxed all day. 2. I/O bound, current SSD can't read/write so many parallel things fast enough. 3. Space bound, 1 TB is clearly not enough.

I've had more "I can't believe it's this good" moments with GPT5.5 than any other model since Opus 4.5. It's shockingly, scarily capable. Days and days of amazing progress. All steering, no handwriting. Yet utterly delightful to conduct its coding. So, so good.

@techgirl1908 clanker wrote the skill, feel free to adapt.

@steipete why does it start like this "Audit Codex/OpenClaw skills"? does it not work for all skills, which are by default harness-agnostic?

@jassdotgg I have all of them on my block list, I see 6 replies. How many are there?

@steipete speaking of purging, how about literally every other reply being an AI reply in this post

@AJalomaki You relax it in the description, not the skill

Bad advice. I have measured evidence. Worst place to relax grammar is in system prompts or equivalent, like skills. x.com/AJalomaki/stat…