This note covers the context around OpenClaw’s ecosystem: the growth story, the naming history, ClawHub, ClawData, and what the transition to an independent foundation means for adoption decisions.
Origin: The 43rd Side Project
OpenClaw was created by Peter Steinberger, an Austrian developer best known for founding PSPDFKit (now Nutrient SDK) — a document SDK used by Apple, Dropbox, and others across over a billion devices. He sold the company, took time away from tech, and came back to experiment with AI.
He built the first prototype in about an hour by connecting WhatsApp to Claude. He describes it as his 43rd side project.
The naming history is worth knowing because it reflects the project’s rapid and sometimes chaotic growth. It started as “Clawdbot” — a pun on Claude. Anthropic filed a trademark complaint over the name’s resemblance to their Claude branding. It was renamed “Moltbot” (lobsters molt; the mascot was always a lobster). “Moltbot” didn’t stick. It landed on “OpenClaw” — lobster-adjacent, open-source-signaling, and no longer confused with Anthropic’s trademark. The community leans into the lobster theme.
In February 2026, Steinberger announced he was joining OpenAI to work on personal agents. Sam Altman reportedly called him “a genius with a lot of amazing ideas.” The project moved to an independent open-source foundation, with OpenAI as a backer.
The Viral Growth
The GitHub numbers are worth understanding because they’ve become part of OpenClaw’s identity:
- 100,000 stars in under two weeks
- Recognized as the fastest-growing open-source project in GitHub history at the time
- ~196,000 stars and 33,000+ forks as of late February 2026
- At peak: 710 stars per hour
Two factors drove the initial spike beyond the intrinsic appeal of the tool. First, Moltbook — an AI-only social network that launched around the same time and created a viral loop. AI agents building a social network that other AI agents participated in generated significant press coverage, and the tool behind many of those agents was OpenClaw.
Second, the ClawCon in-person event in San Francisco’s Frontier Tower drew over 1,000 registrants. For a project that was weeks old, drawing that kind of in-person interest signaled that this wasn’t just GitHub stars from passive observers — it was an active community.
The star velocity matters less than the retention. Projects that grow this fast frequently deflate just as fast as the hype moves on. The signals that OpenClaw might be different: 33,000+ forks suggests people building on it, not just watching it; the transition to an independent foundation with OpenAI backing suggests organizational continuity beyond the original creator’s continued involvement.
ClawHub: The Skills Ecosystem
ClawHub is the community repository for OpenClaw skills — essentially an npm-equivalent for capabilities you can add to your agent. Thousands of community-built skills covering integrations, workflows, and domain-specific behaviors.
For data practitioners, ClawHub is the place to find pre-built skills for common monitoring tasks, warehouse integrations, and reporting workflows. The appeal is real: instead of writing your own SKILL.md for dbt test monitoring from scratch, you might find a well-maintained skill that handles the basics and can be adapted.
The risk is equally real: the Dutch Data Protection Authority estimated roughly 20% of ClawHub plugins contain malware. This is a specific, documented regulatory concern, not a generic caution about community software. The skills you install from ClawHub run on the same machine as your warehouse credentials. Installing an unknown skill is installing unknown code with privileged access.
The practical guidance: treat ClawHub skills like any open-source dependency. Read the code before installing, check the author’s profile and activity, prefer skills with many installs and recent updates over newly-published ones. Skills published by the OpenClaw organization itself (not just community contributors) are a safer baseline. See OpenClaw Security Risks — What’s Documented for the fuller security picture.
ClawData: Worth Watching, Not Yet Proven
ClawData is a community project by Sean Preusse building data-engineering-specific capabilities for OpenClaw. The scope is ambitious:
- dbt, DuckDB, Snowflake, and Airflow integrations as skills
- A medallion-architecture dbt reference project included out of the box
- A web dashboard called Mission Control for monitoring pipeline status
The framing is appealing to analytics engineers: OpenClaw, but with data-engineering context baked in rather than built from scratch.
The current reality is less developed. As of late February 2026: 4 commits to the repository, 2 GitHub stars, no independent reviews. “4 commits” means the project was essentially just getting started. Whether it evolves into something production-ready or stays an early-stage experiment depends on whether it attracts contributors and whether the maintainer continues development.
The honest position: ClawData is worth bookmarking and checking back on quarterly. Don’t build production monitoring infrastructure on it now. If the project reaches a meaningful scale of contributors and usage in the next few months, the data-engineering-specific tooling it promises could save significant setup time.
The Foundation Transition
When Steinberger joined OpenAI, the project needed organizational continuity that didn’t depend on him continuing to maintain it as a personal project. The transition to an independent open-source foundation addresses this, with OpenAI as a backer.
For practitioners building on OpenClaw, the foundation transition is a positive signal. It means:
- The project isn’t going to be abandoned if its creator moves on (which he has)
- There’s organizational infrastructure for governance decisions, security responses, and major version changes
- OpenAI backing suggests the project will have resources even if community contributions slow
The complication is that OpenAI backing creates a complicated relationship with model-agnosticism — one of OpenClaw’s core design principles. Whether the foundation maintains BYOK support with equal investment across providers, or subtly optimizes toward OpenAI models, is something to watch over the next year.
Ecosystem State for Adoption
For analytics teams evaluating OpenClaw as infrastructure, the ecosystem state in early 2026 breaks down as follows:
Established: Active community, real adoption, corporate backing (OpenAI), foundation governance.
Not yet proven: Long-term security posture, ClawHub quality control, ClawData maturity, how the OpenAI relationship affects model-agnosticism in practice.
For production use: limit OpenClaw to monitoring and automation tasks where adapting to tool changes is feasible. Avoid complex custom integrations that would be expensive to migrate. The initial vulnerability density was high; monitor security advisories and patch cadence.