Skip to main content
Glama
base-flash-arb-mcp

base-flash-arb-mcp

README.md
# base-flash-arb-mcp

MCP server for detecting flash loan arbitrage opportunities across DEXes on Base mainnet.

## Tools

| Tool | Description |
|------|-------------|
| `detect_arb_opportunity` | Compare token prices across Uniswap V2, V3, and Aerodrome. Find profitable arb routes. |
| `get_pool_reserves` | Get reserves/liquidity for a token across all known DEX pools on Base. |
| `estimate_flash_profit` | Estimate flash loan profit between two pools after gas costs. |
| `scan_top_tokens` | Scan top traded tokens on Base for arb opportunities via DexScreener. |
| `get_price_across_dexes` | Get current price of a token on all available DEXes. |
| `check_sandwich_risk` | Analyze recent trades for sandwich attack patterns (bot activity). |
| `get_mempool_pending` | Check pending transactions for front-run detection. |

## Install

```bash
npx -y base-flash-arb-mcp
```

## Configuration

Set `RPC_URL` environment variable to override the default Base RPC (`https://mainnet.base.org`).

### Claude Desktop

```json
{
  "mcpServers": {
    "base-flash-arb": {
      "command": "npx",
      "args": ["-y", "base-flash-arb-mcp"],
      "env": {
        "RPC_URL": "https://mainnet.base.org"
      }
    }
  }
}
```

## DEXes Covered

- Uniswap V2 (Factory: `0x8909Dc15e40173Ff4699343b6eB8132c65e18eC6`)
- Uniswap V3 (Fee tiers: 0.05%, 0.3%, 1%)
- Aerodrome (Volatile + Stable pools)

## How It Works

All tools are **read-only** -- no private key needed. The server queries on-chain pool reserves, uses DEX quoters for accurate pricing, and compares across exchanges to find price discrepancies that could be exploited via flash loans.

Flash profit estimation accounts for:
- Aave flash loan fee (0.09%)
- Base gas costs (dynamically estimated)
- Slippage warning

Sandwich detection uses heuristic analysis of Swap events to identify same-block multi-transaction patterns from single addresses.

## Related MCP Servers

| Package | Tools | What it does |
|---------|-------|-------------|
| `obsd-launchpad-mcp` | 14 | Deploy tokens, trade, earn OBSD |
| `base-security-scanner-mcp` | 8 | Scan contracts for vulnerabilities |
| `base-price-oracle-mcp` | 7 | On-chain price feeds from DEX pools |
| `base-multi-wallet-mcp` | 8 | Coordinated multi-wallet trading |
| `base-gasless-deploy-mcp` | 5 | Gasless ERC-20 token deployment |
| `base-flash-arb-mcp` | 7 | Detect arbitrage opportunities |
| `base-token-sniper-mcp` | 5 | Discover & trade new launches |
| `base-wallet-toolkit-mcp` | 7 | Wallet balances, gas, tokens |
| `base-contract-reader-mcp` | 6 | Read any smart contract (free) |
| `create-mcp-server-cli` | - | Scaffold a new MCP server |

## License

MIT

TDQS

A3.5/5.0

Scored across 7 tools

Disambiguation5/5

Each tool has a clearly distinct purpose with no overlap: check_sandwich_risk focuses on attack patterns, detect_arb_opportunity finds profitable routes, estimate_flash_profit calculates profit after gas, get_mempool_pending checks pending transactions, get_pool_reserves retrieves liquidity data, get_price_across_dexes provides price comparisons, and scan_top_tokens scans for opportunities. The descriptions clearly differentiate their functions, making misselection unlikely.

Naming Consistency5/5

All tool names follow a consistent verb_noun pattern with snake_case, such as check_sandwich_risk, detect_arb_opportunity, and get_pool_reserves. This uniformity makes the tool set predictable and easy to navigate, with no deviations in naming conventions across the seven tools.

Tool Count5/5

With 7 tools, the count is well-scoped for the server's purpose of arbitrage and risk analysis on Base. Each tool serves a specific role in the workflow, from detection to estimation and monitoring, without being excessive or insufficient for the domain.

Completeness4/5

The tool set covers core aspects of arbitrage and risk analysis on Base, including opportunity detection, profit estimation, and monitoring tools. However, there is a minor gap in execution tools (e.g., executing trades or managing positions), which agents might need to work around, but the surface is largely complete for analysis and detection purposes.