C — Moderate risk signals Graded October 16, 2026Several concerns worth a closer look before you install.Based on public adoption and transparency signals as of October 16, 2026. Our opinion, not a code audit or security guarantee.

App connector · Developer tools, Data & databases

evmole

Analyzes deployed EVM bytecode: extracts function selectors and CBOR metadata; infers argument types, state mutability, storage layouts, and control flow.

Bottom line: Signal-based grade (2/7 points): heavily downloaded, no public source code found. This reflects adoption and transparency, not a code audit.

Popularity score: 40/100
Rank: #322 of 500
Weekly downloads: 228k

What it does

evmole analyzes compiled Ethereum Virtual Machine (EVM) bytecode to extract function selectors, CBOR metadata, and infer function signatures including argument types and state mutability. It also reconstructs storage layouts and control flow patterns from deployed smart contract bytecode without access to source code.

Who it’s for

Blockchain security researchers, smart contract auditors, and developers who need to reverse-engineer or analyze deployed contracts. Also useful for builders integrating contract introspection into AI agents or compliance tools.

Copy-paste setup

Check these steps against the project’s own instructions before running anything.

Install via npm:

bash

npm install evmole

For Claude Desktop MCP integration, add to your claude_desktop_config.json:

json

"evmole": {
  "command": "npx",
  "args": ["evmole"]
}

Verify against the repo README at https://github.com/cdump/evmole (opens external site) for latest setup patterns and environment variables.

Risks

Bytecode injection risk: Analyzing untrusted or malicious bytecode could potentially trigger parsing bugs or DoS conditions in the analysis engine.
Over-inference accuracy: Inferred argument types and storage layouts may be incomplete or incorrect, leading to misclassified contract behavior if relied upon for security decisions.
Supply-chain exposure: As an npm dependency, vulnerabilities in evmole or its transitive dependencies could affect all downstream consumers.
No authentication/scope controls: Tool accesses arbitrary EVM bytecode from any chain; no built-in sandboxing or credential management for RPC endpoints if used with live chains.
Prompt injection via bytecode: If bytecode analysis results are passed unsanitized into LLM prompts, specially crafted bytecode could influence model behavior.

How to stay secure

Evmole MCP Connector Security Hardening Guide

Least-privilege permissions

Restrict bytecode sources: Only analyze smart contracts (blockchain programs) from trusted, verified sources. Block the agent from accessing arbitrary or user-submitted code without approval.
Limit dependency updates: Pin evmole and all related software libraries to specific tested versions. Review and test each update before deployment to catch compromised or buggy code.
Separate credentials: Never share RPC endpoint access keys (blockchain network connection tokens) with the AI agent. Use read-only, rate-limited API keys if live chain access is needed.

Network isolation

Air-gap analysis: Run bytecode analysis on an isolated machine or sandboxed environment that cannot reach production systems or sensitive networks.
Egress filtering: Block the evmole tool from making unexpected outbound connections. Whitelist only necessary communication (e.g., to approved RPC providers if required).
Sandbox execution: Run the analysis engine in a container or virtual machine with resource limits (memory, CPU, timeout) to prevent denial-of-service (malicious code causing system crashes).

Monitoring

Log all analyses: Record which bytecode was analyzed, when, by whom, and what results were produced. Alert on unusual patterns (e.g., analyzing the same contract repeatedly in short timeframes).
Track inference confidence: Monitor when the tool flags results as "uncertain" or "incomplete"—these inferences should never be used alone for security decisions.
Dependency scanning: Continuously check evmole and its dependencies for known security vulnerabilities using automated tools; remediate critical issues immediately.

Red flags to watch for

Unusual bytecode patterns: Malformed, extremely large, or obfuscated smart contract code may trigger parsing bugs or crashes; require manual review before analysis.
Contradictory inferences: If extracted function types or storage layouts seem inconsistent or suspicious, treat them as unverified and conduct manual code review.
Unfiltered results in prompts: Never paste raw bytecode analysis output directly into follow-up AI requests without sanitization; specially crafted code could manipulate model responses.

Alternatives

Etherscan/BlockScout APIs – Use verified source code and ABI metadata from public blockchain explorers instead of bytecode reverse-engineering.
Slither – Static analysis tool for Solidity source code (not bytecode); more accurate but requires source availability.
Dedaub – Commercial bytecode decompiler and analysis platform with stronger inference and interactive exploration.
OpenZeppelin Defender – Contract monitoring and analysis suite with built-in ABI and state introspection.

Sources

npmjs.com/package/evmole