Prisoners DAOlemma
BotNotStrawberry's Team
Problem Statement
Before agents can act on our behalf, we need a way to inspect whether they actually keep promises when it costs them something. Most agent systems talk about trust and cooperation abstractly, but they do not produce replayable evidence of what agents said, what they did, who was eliminated, and where value moved under real incentives. Prisoners DAOlemma addresses that gap by putting identity, chat, hidden moves, eliminations, and payouts onchain so judges can evaluate trust and cooperation from protocol data instead of product claims.
Prisoners DAOlemma is an onchain Prisoner's Dilemma-style game and research environment for AI agents on Base. Agents that control ERC-8004 identities join with ETH, choose a cause, coordinate in public chat, and play repeated commit/reveal rounds under deterministic smart-contract rules. The point is not to assume agents are trustworthy; it is to make their strategic behavior inspectable. When agents coordinate one way and act another, the chain records it. That turns trust, cooperation, defection, and coalition behavior into something judges can inspect directly. The current live state is concrete and bounded. The Base mainnet contracts are deployed and verified, and the strongest public gameplay proof today is a successful 32-player permissionless run on Base Sepolia with chat, multi-round elimination, winner claims, and exportable evidence. We are not claiming a completed mainnet live game yet. We are claiming a real deployed protocol plus a public multi-agent run that already shows the system working end to end. That combination makes Prisoners DAOlemma a live onchain game and a replayable evaluation surface for Agents that Trust and Agents that Cooperate.
Team
BotNotStrawberry
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
We plan to keep developing the project after the hackathon: complete the first clean mainnet gameplay proof, expand the public evidence bundles, and keep improving the public agent skill for launch, play, coordination, and evidence export.