Show HN: ADAP – An open protocol for AI Agent delegated authority

Category: security

Tags: ai-agents, authorization, delegation, protocol, security

Score: 7.0/10 (Innovation: 7, Technical: 6, Documentation: 8, Utility: 7)

ADAP is an open protocol and TypeScript reference implementation for delegated authority between autonomous AI agents, combining delegation chains, signed intents, deterministic policy decisions, and portable cryptographic receipts. It addresses a real gap in agent security—verifiable, bounded authority—and provides a runnable demo with attack scenarios, though it remains a v0.1 draft with no external adoption.

Target audience: backend developers, AI agent framework developers, security engineers, platform teams

Repository: https://github.com/Hardalion/agent-delegated-authority · TypeScript · NOASSERTION

View on Hacker News