Odei

关于

Odei is a Model Context Protocol (MCP) server that seamlessly integrates AI agents with the ODEI World Model, a sophisticated constitutional knowledge graph. It empowers agents like Claude with critical functionalities, including the ability to query a 91-node knowledge graph across six semantic domains, validate actions against constitutional safety guardrails before execution, obtain instant agent trust scores, and conduct comprehensive EVM smart contract audits using EVMbench methodology. This server acts as a crucial backbone for developing safe, compliant, and context-aware AI applications.

主要功能

• 0 GitHub stars
• Perform EVM smart contract security audits using EVMbench methodology.
• Validate agent actions with guardrail checks (APPROVED/REJECTED/ESCALATE) across 7 constitutional layers.
• Query a 91-node constitutional knowledge graph across 6 semantic domains.
• Receive instant agent trust scores and context signals for high-frequency decisions.
• Easily integrate with MCP clients like Claude via simple configuration.

使用案例

• Integrating AI agents (e.g., Claude) with a rich constitutional knowledge base for context and decision-making.
• Implementing robust AI safety guardrails to validate agent actions before execution in sensitive scenarios.
• Automating security audits for EVM smart contracts to identify vulnerabilities.
• Providing real-time trust scores and contextual signals to AI agents for dynamic adaptation.