Published Papers

Keyword: Multi agent systems ×
1 papers found
Parakletos: On-Chain Identity and Accountability Architecture for Autonomous AI Agents in Trust-Critical Systems
Finance & AI · Vol. 1, No. 2 · Aug 2026
As autonomous AI agents increasingly operate in regulated financial and enterprise environments, the absence of verifiable identity, immutable action provenance, and enforceable authorization presents a critical infrastructure gap. This paper introduces Parakletos, a cryptographic accountability framework deployed on Solana, providing Agent Passports that bind identity, permissions, and action logs on-chain. Using Address Lookup Tables for scalable node management and a hardware-anchored two-node quorum for deterministic validation, Parakletos successfully coordinated 120 autonomous agents in a single session on Solana Devnet. We present the architecture, implementation, and implications for compliance-native multi-agent AI deployment.