sentien-labs/verdictswarm-mcp
🔌 MCP Serversentien-labs
An MCP server utilizing six adversarial AI agents to analyze and verify crypto token safety across multiple chains.
VerdictSwarm-MCP represents a significant advancement in automated security auditing for decentralized finance. Unlike traditional scanners that rely on static heuristics, this MCP server employs a swarm of six specialized AI agents. These agents operate through an adversarial consensus mechanism, where each agent independently evaluates token metadata, smart contract patterns, and on-chain activity. By debating their findings, the swarm minimizes false positives and provides a more nuanced risk assessment.
Key features include real-time detection of common malicious patterns like honeypots and rug pulls, as well as complex behavioral analysis such as holder concentration and wash trading detection. The architecture is built specifically for the Model Context Protocol (MCP), allowing AI assistants and LLM-powered interfaces to query the swarm directly. This enables developers and traders to integrate high-fidelity, multi-agent security insights into their existing workflows, significantly reducing the manual effort required to vet new tokens across Solana, Base, and Ethereum.
💡Highlights
- ├─6-agent adversarial consensus
- ├─Multi-chain: Solana, Base, Ethereum
- └─Detects rug pulls and honeypots
🎯For
- ├─DeFi traders
- ├─Smart contract auditors
- └─Web3 developers