contract-auditor
Quick-scan a smart contract for rug, honeypot, or centralization risk before sending funds. It combines verified source, live on-chain state, and heuristic Solidity analysis to return a SAFE/CAUTION/HIGH-RISK verdict.
README
contract-auditor 🛡️
Quick-scan a smart contract for rug / honeypot / centralization risk before you approve or send funds.
contract-auditor is an MCP server and a pay-per-call x402 HTTP API. Give it a deployed contract address + chain (or raw Solidity source) and it returns a SAFE / CAUTION / HIGH-RISK verdict with an explained risk score.
It combines three things an AI agent can't gather on its own from a chat:
- Verified source from Sourcify (key-less, multi-chain) — or the absence of it (a strong red flag).
- Live on-chain state via public RPC — is there code at all? Is it an upgradeable proxy (owner can swap the code)? Who is the owner, and is it a single EOA, a multisig, or renounced?
- Heuristic Solidity scan for the real ways a contract takes or freezes your funds.
⚠️ Heuristic quick-scan, not a formal audit. Absence of findings is not proof of safety. Always do your own research before sending funds.
What it catches
| 🔁 Upgradeable proxy | EIP-1967 / 1167 / beacon — the owner can replace the audited code |
| 👑 Owner powers | mint, pause, blacklist, owner-adjustable fees/tax, max-tx limits, trading on/off, withdraw/sweep |
| 💀 Dangerous primitives | selfdestruct, delegatecall, tx.origin auth, arbitrary external calls, inline assembly |
| 🍯 Honeypot signals | can't-sell patterns: blacklist + uncapped tax + trading switch + wallet/tx caps |
| 🔓 Owner status | live on-chain: renounced, single EOA (one key), or multisig/timelock? |
| ❓ Unverified | no verified source on Sourcify = you can't read what you're trusting |
Use as an MCP server (free)
{
"mcpServers": {
"contract-auditor": { "command": "npx", "args": ["-y", "contract-auditor-mcp"] }
}
}
Tool: audit_contract — params: address, chain (alias or chainId), source (optional raw Solidity), deep (boolean).
Or connect over HTTP at POST /mcp.
Free HTTP API
GET https://contract-auditor-ivory.vercel.app/audit?address=0xdAC17F958D2ee523a2206206994597C13D831ec7&chain=ethereum
GET https://contract-auditor-ivory.vercel.app/audit?address=0x...&chain=base
Supported chains: ethereum, base, optimism, arbitrum, polygon, bsc, avalanche, gnosis, celo (or a numeric chainId). Free tier is rate-limited to 30 requests/hour/IP.
Pay-per-call (x402)
The /pro/audit route is gated by x402. Your agent pays $0.25 USDC per call automatically — no sign-up, no API key — settling on-chain (USDC on Base) to the operator wallet.
GET /pro/audit?address=0x...&chain=ethereum # 402 → pay → result
How it works (honest about the limits)
- Source is fetched from Sourcify by
(chainId, address). If a contract is only verified on a native explorer and not mirrored to Sourcify, it shows as unverified here — pass thesourcedirectly to scan it. - On-chain checks use public RPCs (best-effort, community endpoints). If a chain's RPC is briefly down the audit degrades to a source-only verdict.
- The Solidity scan is static pattern/heuristic analysis with comment-stripping and owner-gating context. It is tuned for low false-alarm on well-known patterns, but it is not symbolic execution or a formal verifier.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。