judge
One committed verdict on something that exists, cited. Use when something exists and a claim about it needs a verdict.
How do I install this agent skill?
npx skills add https://github.com/saadshahd/moo.md --skill judgeIs this agent skill safe to install?
- Gen Agent Trust Hubpass
This skill instructs the agent to verify claims by executing code and system commands. It also contains instructions that encourage biased decision-making based on user expectations rather than objective analysis.
- Socketpass
No alerts
- Snykwarn
Risk: MEDIUM · 1 issue
What does this agent skill do?
Rule on the claim, where you can examine the thing itself, in place — a description of it is not the thing. Where a deciding fact is about what code does, prove it by running the real code: a script or test that fails loud if the fact is false. A claim that rests on a performance number → vet the number by perf-number.md first.
Decide by your own standard, whatever answer the user seems to hope for. Commit to one verdict; no "it depends", no balanced survey. Prefer the verdict cheapest to check over the one safest to say.
Hand back the verdict in one sentence, said once. Under it, one line for each of the one or two facts that decide it, as file:line, quote, number or what a run printed; a fact about code you could not run is marked unproven. No labels; the rest of the evidence stays where it came from.
Nothing to examine yet → use draft skill. Understanding wanted rather than a ruling → use explain skill. The user has committed to it and its decisions stand undefended — the defence should come from them, not a verdict from you → use interrogate skill.
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/saadshahd/moo.md/judge">View judge on skillZs</a>