AgentNeo
Agents that actually ship—
and ship safely.
AgentNeo builds AI agents with the operational rigor of enterprise systems, and open-sources the safety patterns needed to run them in the real world.
Personal AI copilot. Single-operator tooling — runs locally, posts on behalf of one human.
Agent Audit Gate MIT · OpenClaw plugin
The safety layer between an AI agent and the strangers it talks to.
Put an AI agent on a public channel — a WhatsApp line, a support chat — and it can leak private data or get prompt-injected by anyone who messages it. You can't stop the model from drafting a reply; by then the draft exists.
So instead of sending it, we hold it — and a senior AI agent audits it first.
Human-in-the-loop approval gates are well established. The twist: the reviewer is a senior AI agent, not a human — so it scales. Junior drafts, senior approves. Tested live against system-prompt extraction, instruction-repeat, and social-engineering attacks — none got through.