Stop Installing Things You Don't Need

The problem

The OpenClaw ecosystem pours out new tools, extensions, skills every week. Most of them look useful. Most of them aren’t, for me, right now.

My default was install-and-hope. Try it, see if it sticks, realise I already had something that does the same thing. Or worse, install it, forget about it, and wonder six weeks later why my workspace feels cluttered.

The real cost wasn’t the install. It was the cognitive load of trying, comparing, and figuring out afterwards whether any of it was worth it.

What it does

Instead of trying each new thing, this pack checks what I already have first and decides if the new tool adds anything.

Three skills, one question each:

Should I install this? Four checks before anything touches the system. What it actually does, whether it solves a real problem, whether something already covers it, and whether the repo is even alive.

Is this external code safe? Skills run inside the agent with broad access. They get reviewed for red flags before they run.

Then after install: did it actually work? Three tiers of validation. Won’t say “done” until it’s confirmed.

Say “install X” and the pipeline runs. Or run the auditor on its own against anything already installed. It won’t try to reinstall, just verify.

What’s fun about it

The system tells me “shiny toy, skip” or “real gap, install” and I move on either way without the overhead. No more wasted afternoons.

The skill-vetter came from adamb0mbNZ on the OpenClaw subreddit. I cleaned it up and bundled it with the two I built.

Get it

Good tools get better inside a system.