drainbrain-mcp-server

DrainBrain MCP Server

MCP server for DrainBrain — AI-powered Solana token rug pull detection.

Gives any MCP-compatible AI agent (Claude, Copilot, Cursor, custom agents) the ability to scan Solana tokens for rug pull risk using DrainBrain's 4-model ML ensemble, trained on 175K+ labeled tokens.

Built by RugSlayer — the Solana pre-crime scanner.

Quick Start

Claude Desktop

Add to your claude_desktop_config.json:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "drainbrain": {
      "command": "npx",
      "args": ["-y", "drainbrain-mcp-server"],
      "env": {
        "DRAINBRAIN_API_KEY": "your-api-key"
      }
    }
  }
}

Claude Code (CLI)

claude mcp add drainbrain -- npx -y drainbrain-mcp-server

Then set your API key:

export DRAINBRAIN_API_KEY=your-api-key

Any MCP Client

DRAINBRAIN_API_KEY=your-key npx drainbrain-mcp-server

Tools

scan_token

Scan a Solana token for rug pull risk. Returns a comprehensive risk assessment including score, honeypot detection, risk flags, and temporal behavioral analysis.

Parameters:

Name	Type	Required	Description
mint	string	Yes	Solana token mint address (base58 public key)

Returns:

• Risk score (0-100) with severity classification
• Risk level (safe / low / medium / high / critical)
• Rug detection status and rug stage (1-5 progression)
• Honeypot analysis with buy/sell tax percentages
• Score breakdown by model component
• Risk flags (specific issues detected)
• Temporal analysis (behavioral pattern prediction with confidence and estimated time to pull)

Example prompt:

"Scan this Solana token for rug pull risk: So11111111111111111111111111111111111111112"

Example output:

Token: So11111111111111111111111111111111111111112
Risk Score: 12/100 SAFE
Risk Level: low
Rug Detected: NO
Honeypot: No

Score Breakdown:
  heuristic: 15
  ml_v1: 8
  ml_v2: 10
  temporal: 14

batch_scan

Scan multiple Solana tokens in one call. Returns a summary table with risk scores for each token. Maximum 10 tokens per batch.

Parameters:

Name	Type	Required	Description
mints	string[]	Yes	Array of Solana token mint addresses (1-10)

Example prompt:

"Batch scan these 3 tokens for rug pull risk: Mint1, Mint2, Mint3"

Example output:

Batch Scan Results (3 tokens):

So111111...1112: 12/100 SAFE | low | Rug: NO
EPjFWdd5...Dt1v: 8/100 SAFE | low | Rug: NO
DezXAZ8z...pump: 85/100 CRITICAL | critical | Rug: YES

health_check

Check if the DrainBrain API is operational and which ML models are available. Use this to verify the service is working before scanning tokens.

Parameters: None

Example prompt:

"Check if DrainBrain is online"

Example output:

Status: healthy
API Version: v1

Models:
  ml_v1: OK v1.0.0 (0.06ms)
  ml_v2: OK v2.0.0 (0.08ms)
  temporal: OK v2.0.0 (0.12ms)
  heuristic: OK v1.0.0

compare_rugcheck

Compare DrainBrain's ML-based score against RugCheck's heuristic-only score for the same token. Shows where ML analysis provides an advantage over traditional approaches.

Parameters:

Name	Type	Required	Description
mint	string	Yes	Solana token mint address (base58 public key)

Example prompt:

"Compare DrainBrain vs RugCheck for this token: DezXAZ8z7PnrnRJjz3wXBoRgixCa6xjnB7YaB1pPB263"

Example output:

Comparison for: DezXAZ8z7PnrnRJjz3wXBoRgixCa6xjnB7YaB1pPB263

DrainBrain (ML Ensemble):
  Score: 85/100 CRITICAL
  Risk Level: critical
  Method: ensemble-temporal

RugCheck (Heuristic):
  Trust Score: 45
  Risk Level: medium

Analysis:
  Agreement: No
  Note: DrainBrain detected temporal rug patterns not visible to heuristic analysis

Risk Score Guide

Score	Level	Meaning
0-19	SAFE	Low risk - no rug indicators detected
20-39	LOW	Minor concerns - exercise normal caution
40-59	MEDIUM	Moderate risk - several warning signs present
60-79	HIGH	High risk - significant rug pull indicators
80-100	CRITICAL	Extreme risk - strong rug pull signals detected

Configuration

Variable	Required	Default	Description
DRAINBRAIN_API_KEY	Yes	-	Your DrainBrain API key
DRAINBRAIN_API_URL	No	https://rugslayer.com	API base URL

Get an API Key

• Visit rugslayer.com/drainbrain
• Sign up and generate a key
• Free tier includes 100 scans/day

How DrainBrain Works

DrainBrain uses a 4-model ML ensemble trained on 175K+ labeled Solana tokens to predict rug pulls before they happen. The system analyzes liquidity patterns, on-chain behavior, temporal signals, and token authority configuration to produce a single risk score (0-100) with 5-stage rug progression tracking.

Unlike heuristic-only tools, DrainBrain detects behavioral patterns that precede rug pulls - often identifying threats before any price movement occurs.

Requirements

• Node.js 18+
• DrainBrain API key (get one at rugslayer.com/drainbrain)

Links

• RugSlayer - Full token scanner platform
• DrainBrain - API documentation
• Chrome Extension - Browser-based scanning
• GitHub - Source code

License

MIT