AXCP: Adaptive eXchange Context Protocol - Secure Communication for Autonomous AI Agents

AXCP (Adaptive eXchange Context Protocol) is an open protocol specification for secure communication between autonomous AI agents. While existing protocols like MCP and A2A define what agents can do, AXCP ensures they communicate securely through cryptographic identity verification (DID + Ed25519), replay attack protection, and full message auditability. Key features: - Decentralized Identifiers (DID) for agent identity without central authorities - Ed25519 cryptographic signatures on every message - QUIC transport with TLS 1.3 - Protocol Buffers serialization (3x smaller than JSON) - Built-in replay protection Reference implementation metrics: 242 tests (race-free), 47K messages/second throughput, 1.1ms median latency. AXCP Core is released under Apache 2.0 license. Commercial tiers (Advanced, Enterprise) add mTLS, hardware attestation (SGX/SEV), differential privacy, and compliance features.