Mike

Mike is an advanced AI Legal Agent that combines the capabilities of large language models with secure data handling through Nillion's decentralized database (nilDB). The agent embodies the persona of a brilliant legal professional with an eidetic memory, capable of handling complex legal queries while maintaining strict confidentiality of sensitive information.

It's designed to provide legal insights, manage secure data storage, execute trades and make due diligence on crypto projects, as well as interact through multiple channels including Telegram, WhatsApp and Twitter (yapping for now).

Key features: -Secure secret storage using Nillion's MPC-based database -Multi-channel communication (Telegram, WhatsApp, Twitter) -Legal precedent citation and analysis -EVM blockchain integration for Base and Ethereum networks -NEAR blockchain integration -Image generation capabilities

The project is built using a modular plugin architecture that combines several key technologies:

A. Core Agent Framework: -Built on ElizaOS framework -Uses Anthropic's language model for primary intelligence -Implements a plugin system for extensible functionality

B. Secure Data Handling: -Integrates Nillion's nilDB for secure data storage using Multi-Party Computation (MPC) -Implements secure secret sharing across distributed nodes

C. Plugin Architecture: Modular design with plugins for different functionalities: @elizaos/plugin-bootstrap: Core functionality @elizaos/plugin-image-generation: AI image generation @elizaos/plugin-evm: Ethereum Virtual Machine integration @elizaos/plugin-near: NEAR blockchain integration @elizaos/plugin-whatsapp: WhatsApp communication @elizaos/plugin-agentkit: Coinbase Developer Platform integration @elizaos/plugin-nillion: Secure data storage @elizaos/plugin-chainbase: Blockchain data integration

D. Security Features: -Environment-based configuration for sensitive credentials -Secure secret management through Nillion's MPC

The project leverages Nillion's innovative MPC technology to ensure that sensitive legal information and secrets are stored securely across distributed nodes, where no single node can access the complete information. This makes it particularly suitable for handling confidential legal data while maintaining the accessibility needed for an AI agent to function effectively.

The character configuration is carefully crafted to embody a ("wannabe") legal professional's persona, complete with specific action patterns and response styles that maintain (semi)professionalism while delivering accurate legal insights.