# Better Agent Auth — Bereket Habtemeskel & Paola Estefania, Better Auth

## Executive summary

This workshop details the Agent Auth protocol, a solution designed to secure autonomous AI agents operating services for users or organizations. The core principle shifts from granting broad user credentials (acting *as* the user) to giving specific, traceable authority (acting *for* the user). Key mechanisms include agent identity management using private keys, fine-grained capability discovery via directories, and granular authorization that limits actions to explicitly granted capabilities.

## Key takeaways

- Shift from Credentials to Authority: Instead of providing agents with full user credentials, the protocol treats the agent as a principal actor with its own distinct authority. This is analogous to hiring an assistant rather than giving them CEO access.
- Capability Discovery and Authorization: The system requires a directory mechanism for agents to discover what services are available and what specific actions (capabilities) they can perform, moving beyond broad scopes like 'read.'
- Agent Identity and Traceability: Every agent must possess a unique identity (private key). This allows the system to maintain comprehensive audit logs, track exactly which agent performed an action on behalf of which user, enabling immediate revocation if misuse occurs.

## Technical details

- Agent Auth Protocol Design: The protocol addresses three main issues: 1) Agent discoverability (how agents find services/capabilities), 2) Fine-grained authorization (limiting actions to specific capabilities, not broad scopes), and 3) Identity management.
- Capability Mapping: Capabilities are more granular than traditional OAuth scopes. The system aims to map an agent's intent directly to a defined tool or set of tools/capabilities. For services lacking native support, the protocol suggests using OpenAPI specifications (JSON) for capability extraction.
- Identity and Security Flow: Agents are assigned unique identities with private keys, allowing them to sign tokens. This changes the paradigm from viewing the agent as hiding behind the user to treating the agent as a principal actor. The architecture involves an SDK (for key generation), a Server side (for policy verification/authorization grants), and a Directory layer.
- Authorization Enforcement: The system enforces capabilities at runtime, requiring explicit user approval for actions beyond basic read access. Revocation of the agent's identity immediately prevents further unauthorized activity.

## Practical implications

- For developers building AI applications, this protocol mandates implementing robust identity and authorization layers rather than relying on simple credential passing.
- It provides a clear architectural blueprint for integrating autonomous agents into enterprise systems while maintaining strict security boundaries and auditability.
- The focus on capability discovery allows for the secure integration of legacy or non-standard services by leveraging existing standards like OpenAPI.

## Topics

AI Agents, Authentication Protocols (Auth), OAuth/OIDC, Capability-Based Security, Identity Management, Build Engineering, Agent Auth Protocol, Better Auth

Source: https://www.youtube.com/watch?v=JvKO40CFq-s