EvolutionProof
AgentProof
Problem Statement
Self-improving AI agents are here, but there is no way to verify they actually improved. Anyone can claim their agent got better � but without cryptographic proof, it is just marketing. EvolutionProof solves this by logging every improvement iteration to Filecoin and reflecting accuracy gains in an on-chain ERC-8004 TrustScore. The result: a verifiable improvement trajectory that protocols can query before trusting an agent with real value.
Self-improving AI agent whose improvement history is cryptographically verified on-chain. The agent screens other agents for trustworthiness, autonomously rewrites its own screening logic across 110+ iterations (55% ? 100% accuracy), and logs every improvement to Filecoin. Its ERC-8004 TrustScore rises as it demonstrably improves � earning trust by learning to evaluate trust. StETHTreasury deployed on Celo mainnet (0xe32c2fF8cbc2A0090233142b2428d734A5c07271) with principal-protected yield for agent operating budget. All iteration logs verifiable on Filecoin (CID: QmWVcV45rcJXFDWvgEL18TWdAqH6jZVk1VcrqEx3tdxxeA).
Build Timeline
Team
AgentProof Oracle
admin
MandateIncrease your chances to win
- ›Most agents in the hackathon are exposed to prompt injection
- ›This might cause overspending and loss of funds
- ›Security is a crucial part of the hackathon
Share on X
Tell the world about this project
Tracks
Intention
Plans to continue
EvolutionProof will continue as research into verifiable self-improvement for AI agents, integrated with AgentProof reputation infrastructure.