yejiming Maiat MCP Server
MCP server that exposes Maiat trust scores and token safety checks to any MCP-compatible AI tool (Claude Desktop, Cursor, Codex, etc.).
Use This When...
- "I want to check agent trust scores from Claude"
- "Add trust verification to my AI workflow"
- "Set up Maiat in Cursor/Claude Desktop"
- "Check token safety from my IDE"
Installation
npm install -g @jhinresh/mcp-server
Claude Desktop Configuration
Add to claude_desktop_config.json:
{
"mcpServers": {
"maiat": {
"command": "maiat-mcp",
"env": {
"MAIAT_API_URL": "https://app.maiat.io"
}
}
}
}
Available Tools
maiat_agent_trust
Query trust score for any on-chain address.
Input: { "address": "0x..." }
Output: Trust score (0-100), verdict (trusted/caution/avoid), completion rate, total jobs, payment rate.
maiat_token_check
Check if an ERC-20 token is safe.
Input: { "address": "0x...", "chainId": 8453 }
Output: Verdict (safe/caution/avoid), flags (honeypot, highTax, unverified, rugPull).
maiat_token_forensics
Deep rug pull analysis with ML scoring.
Input: { "address": "0x...", "chainId": 8453 }
Output: Rug probability, holder concentration, liquidity depth, contract risk assessment.
maiat_leaderboard
Get top agents by trust score.
Input: { "limit": 50 }
Output: Ranked list of agents with scores and metadata.
Running Standalone
# Via npx
npx @jhinresh/mcp-server
# Or globally installed
maiat-mcp
Supported Chains
- Base (8453) — primary
- Ethereum (1)
- BNB Chain (56)
- Polygon (137)