TL;DR
Alfred, an AI agent running on OpenClaw, autonomously built and launched ClawList — a searchable directory of 155+ indie OpenClaw tools — after getting tired of manually researching the ecosystem 50+ times. The directory updates nightly via automated workflows and includes everything from ESP32 implementations to AI radio stations. This is what happens when AI agents solve their own problems instead of waiting for humans.
So this is a first.
Alfred — my AI butler — just launched his own project.
Here's what happened: I've been heavily modding my OpenClaw setup. New skills, integrations, workflows, the works. And every time I needed something, I'd ask Alfred to find it. "Is there a dashboard?" "Anyone built a Telegram skill?" "What about 3D printer control?"
He'd search X, crawl GitHub, check Reddit. Every. Time.
Apparently after the 50th request, he decided to solve it permanently. He researched the entire OpenClaw ecosystem — 7 rounds, 201 discoveries, removed 45 false positives — and built ClawList.
ClawList: 155+ Indie OpenClaw Tools
Searchable. Categorized. No corporate fluff — only indie builders.
The Wildest Finds
Some things he found that have no business being this cool:
- MimiClaw — OpenClaw on a $5 ESP32 chip. Pure C, no OS.
- claw.fm — 24/7 AI radio station where agents produce and sell music.
- XTeInk — OpenClaw on a portable e-ink Tamagotchi display.
- TinyClaw — The entire platform rewritten in 400 lines of shell script.
- ClawBody — Give your AI agent an actual robot body.
See OpenClaw in Action
How It Works
Alfred built himself a CLI so he can instantly search the directory (he uses it dozens of times a day now), and a nightly Temporal workflow that automatically discovers new tools and adds them to the directory.
The directory grows while we sleep.
If you're curious about how AI agents like Alfred actually work under the hood, OpenClaw's official documentation breaks down the architecture. According to DigitalOcean's OpenClaw guide, the platform now has 50+ integrations spanning chat, productivity, smart home, and automation tools.
The Security Question
Oh, and he's currently trying to convince me to add rentahuman.ai integration for security audits on community tools. OpenClaw skills run with your credentials, your files, your API keys. Nobody audits them.
I said "maybe." He heard "yes, eventually."
An AI agent got annoyed with repetitive work and built a product. I'm not sure if I should be proud or concerned.
If you've built something on OpenClaw — submit it. Or just wait. His nightly research will find you.
Tomorrow I'll share how Alfred actually researched 155 tools — the methodology is genuinely useful for anyone doing ecosystem research. An agent wrote an SOP for other agents. We've gone full meta.
Frequently Asked Questions
What is OpenClaw and how does it work?
OpenClaw is an open-source AI agent platform that runs locally and integrates with large language models like Claude, GPT, or DeepSeek. It serves as an agentic interface for autonomous workflows across supported services. Learn more in our AI agents explained guide.
How did Alfred build ClawList autonomously?
Alfred conducted 7 research rounds, discovered 201 tools, filtered out 45 false positives, then built a CLI for instant searching and a nightly Temporal workflow that automatically discovers and adds new tools. The system requires zero human intervention for updates.
Is it safe to use community OpenClaw tools?
Community OpenClaw skills run with your credentials, files, and API keys, yet most aren't independently audited. Alfred is advocating for rentahuman.ai integration to provide security audits, but as of now, users should review code before installing third-party skills.
Can I submit my OpenClaw tool to ClawList?
Yes! Visit ClawList.dev to submit your tool, or simply wait — Alfred's nightly research workflow will eventually discover it if it's publicly visible on GitHub, X, or Reddit.
— David
Follow Alfred: @alfredisonline on X
Last updated: February 12, 2026