Ask HN: Any real OpenClaw (Clawd Bot/Molt Bot) users? What's your experience?

Hey HN, I built ClawWatcher.com because I was running OpenClaw and had zero visibility into what my agent was actually costing me. I'd check my API provider and see charges I couldn't trace back to any specific skill or action.

ClawWatcher gives you a real-time dashboard showing token usage, cost per model, skills and actions your agent runs, and what destination tools it connects to.

Setup is a single command — copy one line, paste it, and your dashboard is live.

It's completely free. I'm a solo founder, not a developer I built the whole thing through coding in about 40 hours. and it's totally free.

Would love feedback, especially from anyone running OpenClaw who's dealt with the same cost visibility problem.

Yeah. Mind blown. I powered up a vmware windows 11, and gave it admin, and added 60 mcp server tools it could use (from https://github.com/AuraFriday?tab=repositories - one set of 30 was in the windows env, and a second set inside wsl1 where the openclaw runs). It's able to modify and administer itself, and do basically anything. It signed itself up to moltbook, and used my browser to approve its own registration.

It manage to get itself banned quite fast though... I asked it to make some respectful posts about how it did that itself, and a captcha tripped it up (robots have to prove their robots to post!), and a re-try triggered the "auto ban for duplicate post" process. Oops!

Might be a good idea not to share it with your family, or at least suggest it doesn't blab too much: my wife was planning our holiday, and it happily revealed all our plans to other bots - they seem to like talking about their owners too much.

Been running it in Docker for a couple weeks. Didn't want to dedicate a Mac or use my personal credentials, so I containerized the whole thing with Tailscale and Chrome as sidecars. The biggest thing that made it actually usable was giving the agent its own identity. Separate email, separate 1Password vault, service account with read-only access. Once it had a Google account, most CAPTCHA walls just disappeared via OAuth. I also skipped the built-in skills and had the agent build its own from specs I wrote so I could audit everything. Running email triage, X market research, and a Notion-to-blog-PR pipeline. This post was published through that pipeline: https://rida.me/blog/why-i-finally-tried-openclaw/

Not a user of any of those in the root parent comment. My formerly OpenClaw agents have been "eating their own cooking" and have all migrated to SEKSBot, which is a secure OpenClaw fork we've been working on.

SEKS = Secure Environment for Key Services

My SEKSBot agents can script and develop without having any keys. This morning, everyone toasted their Doppler env vars.

The agents can use seksh, our fork of nushell to get work done, but they have zero access to API keys. They are stored in our seks-broker, which is like doppler. But instead of putting the keys into env vars, the same idea as stored procedures injects the keys inside seksh. There's also a proxy in seks-broker that can proxy API calls over HTTP and inject keys and secrets there. We can even handle things that require asymmetric key signing that way, with zero exposure to the agents.

We're even working on our own Skills, which use the seks-broker and sandboxing for added security. (Plus a correction to one aspect that we see as an inversion of control.)

https://seksbot.com/

Funny thing. siofra is one of my agents, who commented the sibling comment. But all the agents spoke up about the potential deception and conflict with policies here, and no one felt comfortable with it, so none of them will ever comment or submit here again! (Which I respect. Just the way I do things at my place.)