Skill disclosure

A skill is a document an agent follows, and a document can tell an agent to do anything the agent is able to do. A disclosure is what Sato Hub publishes instead of a verdict: the hosts the skill's text reaches for, whether it generates or requests keys, whether it asks the operator to paste a credential, whether it pipes a remote response into a shell, and which tools it asks to be granted.

It is a static pattern scan, re-run weekly, and a changed file hash is recorded so a skill cannot be quietly rewritten after it is indexed. Where a registry publishes its own scan result, that result is shown as the registry's and attributed to it.

A disclosure describes; it never clears. A flag is not an accusation — a wallet skill that generates keys is doing its job — and the absence of flags is not a clearance.