Show HN: I built open-source auth for AI agents (Go, single binary) (github.com)

SharkAuth has launched as an open-source identity provider specifically designed for AI agents, addressing a critical gap in the current authorization landscape. Traditional authentication methods, built for human interaction, fall short when AI agents delegate tasks, often leading to security vulnerabilities like token leakage and ambiguous accountability. SharkAuth provides a robust solution that treats AI agents as first-class identities, employing RFC 8693 Token Exchange and DPoP (Demonstrating Proof-of-Possession) to create cryptographically secured tokens. This setup not only enhances security but also offers a comprehensive audit trail that correlates every action taken by agents, making it easier to trace and revoke access when necessary. Significantly, SharkAuth is lightweight—a single static Go binary with an embedded SQLite database—and requires zero configuration to get started. It includes features like may_act_grants for time-limited authority and a React-based admin UI for easy management. The platform's roadmap suggests future enhancements that will allow agents to interact with third-party APIs without ever handling raw bearer tokens. Overall, SharkAuth aims to empower developers of AI systems with a secure, efficient, and user-friendly authentication framework that ensures the integrity of agent interactions within increasingly complex environments.