Merged PR author at openclaw/openclaw

I read the source code of the tool I sell, and ship fixes back to it.

Around 41,000 people have opened a pull request against openclaw/openclaw. Of those attempts, only ~6,900 PRs have ever merged into core (the main branch of openclaw/openclaw), roughly a 1-in-6 hit rate against one of the strictest contribution review pipelines in open source. Mine is one of them, reviewed and merged by the project's creator, Peter Steinberger, himself. Every claim on this page is linked back to GitHub so you can verify it yourself.

367K
GitHub stars on the project
~41K
People who opened a PR
~6.9K
PRs that ever merged into core
Mine
Merged by Peter Steinberger
Merged into coreMay 2026·Merged by Peter Steinberger, creator of OpenClaw

I caught and patched a $20-30 per minute cost-runaway bug in OpenClaw core.

openclaw/openclaw#76345 · closes #76293

The bug, in plain English

When an OpenClaw agent talks to a model like Claude Sonnet and the connection silently stalls, the system is supposed to time out and try again. But the retry would hit the same wedged connection, time out again, retry, time out, forever. Every retry is a paid API call.

One reporter logged 761 to 1,384 paid Claude Sonnet 4.6 calls in 60 seconds across two real incidents. $20-30 burned in a single minute, masked by auto-recharge on the provider account so users only found out at billing.

The fix

A circuit breaker at the outer agent run loop. After 5 stalled attempts in a row with no model output, the system stops trying and refuses further calls until the connection actually responds again. Worst case is now roughly $0.10-$0.30 per incident instead of $20-30. Routing semantics intentionally unchanged so existing deployments keep working.

Files touched

  • ·src/agents/pi-embedded-runner/run.ts
  • ·src/agents/pi-embedded-runner/run/idle-timeout-breaker.ts
  • ·src/agents/pi-embedded-runner/run/idle-timeout-breaker.test.ts
  • ·CHANGELOG.md
View the merged PR on GitHubRead the original bug report

Why I think this matters for the work I do.

OpenClaw is one of the fastest-growing open source projects of the year, 367K GitHub stars at the time of writing. The contribution bar is deliberately strict. Around 41,000 people have opened a pull request and only ~6,900 have ever merged, roughly a 1-in-6 hit rate. Refactor-only PRs are auto-rejected, most feature requests get pushed to third-party plugins, and each contributor is capped at 10 open PRs at a time. Getting code into core takes reading the runtime carefully, writing a fix the maintainers agree is the right shape, and surviving a multi-pass AI review pipeline before a human even looks at it.

I think the right consultant for serious OpenClaw work is one who has actually been inside the runtime, not just the docs. When a production incident hits at 2am, the difference between a consultant who watched a YouTube walkthrough and one who has shipped to the file the bug is in is the difference between a long Slack thread and a fix.

I am not the only person who can do this work, and I am not claiming to be the most experienced contributor on the project. What I can say is that the merged PR above is verifiable evidence I do read the source, I do find the bugs, and I do ship the fix when it is warranted. If that is the bar you want from a consultant, we are probably a good fit.

Verify any of this yourself.

Every claim on this page is checkable in public. The work is on GitHub under your own account if you want to log in, and visible without an account either way.

If you want this depth on your OpenClaw build, talk to me.

I take a small number of clients each quarter. Application is the same link below. I read every one personally and tell you in the first reply if we are not the right fit.

Apply for a buildMore about me →