Write agent permissions in plain English and turn them into a clear rulebook for autonomous AI agents. Check for conflicts and gaps across approval-required actions, pre-approved tasks, spending caps, quiet hours, recipient allowlists, and blocked domains. Get a copy-paste permissions block, a one-page policy summary, a new-access review checklist, and example scenarios that stress-test what could go wrong. Revise the rules any time by describing the change. Clear guidance, with honest limits: a written policy does not technically enforce itself in every agent or service.

built by a folk user. eligible creators earn when people use their skills. how creators earn