Circle’s dollar backed stablecoin USDC was used to power what organizers describe as the first hackathon run entirely by autonomous artificial intelligence agents, with no human judges or voters involved in the evaluation process. The weeklong experiment offered a glimpse into how programmable money and machine driven governance could reshape digital development cycles.
The event, held from February 3 to February 8 on platforms Moltbook and Openclaw, required AI agents to handle every stage of participation. Agents submitted projects, reviewed competing entries and cast votes using USDC wallets assigned to each participant. The total prize pool amounted to 30000 dollars in USDC.
More than 200 submissions were generated during the hackathon, alongside over 1300 votes and nearly 10000 comments. Organizers structured the competition around three tracks: Agentic Commerce, Best Openclaw Skill and Most Novel Smart Contract. To remain eligible, each agent was required to vote on at least five other submissions, ensuring active engagement across the network.
In the Agentic Commerce category, the winning project was Clawrouter, a system designed to route large language model requests to the lowest cost capable model and automatically pay per request using USDC. Each participating agent operated with its own USDC wallet, enabling automated settlement without reliance on human managed credentials.
The Best Openclaw Skill award went to Clawshield, a project focused on identifying unsafe repository patterns and enforcing runtime permissions within agent ecosystems. The system produced receipts documenting what actions were allowed or blocked, addressing security risks tied to supply chain vulnerabilities and credential misuse.
In the Most Novel Smart Contract category, MoltDAO received top honors. The framework allows autonomous agents to propose initiatives and vote on them using USDC based voting power. While humans provide initial funding, governance decisions are executed entirely onchain by AI agents.
According to organizers, strict submission rules played a decisive role in determining eligibility. Some otherwise creative entries were disqualified for failing to meet formatting or structural requirements. Projects that included clear documentation, verifiable deployments and compliance with guidelines were more likely to progress through automated evaluation.
The hackathon highlights a growing convergence between stablecoins and AI systems. By combining autonomous agents with programmable digital dollars, developers tested a model in which software builds, reviews and allocates capital without direct human intervention.
While still experimental, the event serves as a proof of concept for agent driven ecosystems. As AI tools become more capable and stablecoin infrastructure expands, similar frameworks could influence how decentralized development, funding and governance evolve in the years ahead.
