Yes or nAI

Introducing our cutting-edge AI Agent bot, engineered for the Eth Global hackathon Agentic, which is set to revolutionize DAO governance. This autonomous agent is designed to streamline decision-making processes by not only submitting proposals and casting votes on demand but also by independently generating well-reasoned rationales for its actions. Leveraging advanced natural language processing and state-of-the-art reasoning algorithms, the bot evaluates complex governance scenarios, drafts insightful proposals, and articulates clear, data-driven justifications for its voting decisions. Its ability to operate autonomously while maintaining transparency and accountability positions it as a pivotal tool in enhancing efficiency and democratization within decentralized autonomous organizations.

Our project is built as a modular, TypeScript-based Node.js application that brings together a suite of interconnected components to enable autonomous DAO governance. At its core, we leveraged the @elizaos/core framework, which provided the foundational abstractions for caching, state management, and action handling. We integrated with external services using partner technologies such as the Twitter client from @elizaos/client-twitter and Ethereum libraries like viem for encoding smart contract function calls and sending transactions.

The project is broken down into several key modules: a caching layer that abstracts between Postgres and SQLite (using adapters like better-sqlite3 for lightweight storage), a command-line chat interface built with Node’s native readline module for testing and interaction, and a suite of governance actions (propose, queue, execute, vote) that interact with DAO smart contracts. Each governance action uses the ABI of the governor contract to encode transactions, and our custom wallet provider handles the nuances of sending transactions to the blockchain. We even built in some robust features like rate limiting and exponential backoff in our digest action using axios—a hacky but effective way to avoid API rate limits when fetching external data.

By piecing these technologies together, we achieved a system that can autonomously generate proposals, digest external data, and participate in DAO governance with minimal human intervention. The modular design allowed us to easily swap out components (for instance, switching between database backends or adding new client integrations) and provided a clear separation of concerns, which was particularly beneficial in the fast-paced hackathon environment. Overall, the combination of partner libraries and some creative engineering (like recursive chat loops and adaptive caching mechanisms) helped us rapidly prototype and iterate on a highly functional and autonomous governance agent.