web3-solidity-audit-mcp

Fail

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The installation instructions suggest fetching and executing shell scripts directly from external sources using curl | bash for Foundry (Paradigm) and Aderyn (Cyfrin).
  • [COMMAND_EXECUTION]: The recommended GitHub Actions CI/CD configuration employs sudo to install binaries into system-protected directories (/usr/local/bin).
  • [EXTERNAL_DOWNLOADS]: The skill depends on several third-party packages and binaries from multiple registries, including NPM (solidity-audit-mcp), PyPI (slither-analyzer, solc-select, halmos), and Homebrew (echidna).
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection where malicious instructions could be embedded in the comments or strings of the Solidity contracts being audited.
  • Ingestion points: Solidity source files provided to the analyze_contract and check_vulnerabilities tools via the contractPath parameter.
  • Boundary markers: The skill does not define specific delimiters to separate untrusted contract code from instructions when presenting findings to the agent.
  • Capability inventory: The MCP server executes sub-processes (Slither, Aderyn, Forge tests) and has file system access to read project directories and write audit reports.
  • Sanitization: No evidence of sanitization or filtering of comments or metadata from audited files is present, which may allow embedded instructions to influence the agent's summary or remediation advice.
Recommendations
  • HIGH: Downloads and executes remote code from: https://foundry.paradigm.xyz, https://raw.githubusercontent.com/Cyfrin/aderyn/dev/cyfrinup/install - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 17, 2026, 04:28 PM
Security Audit — agent-trust-hub — web3-solidity-audit-mcp