AgentHub — Secure AI Agent Access Layer

AgentHub solves secure access for AI coding assistants: policy-controlled access to databases, APIs, and infrastructure. It enforces least-privilege, maintains audit trails, and routes approvals to humans for high-risk operations. Designed for AI-native development where assistants need production access without losing control.

Stack: TypeScript, Node.js, PostgreSQL, Vercel. Founder-built personal project with live demo at agenthub-lyart.vercel.app and GitHub at github.com/theweb3wizard/AgentHub. Related service: AI Agent & Solana Engineering.