firecrawl-mcp

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill's Python scripts are consistent, well-documented wrappers that forward payloads to the OneKey Gateway API. No malicious patterns such as obfuscation, persistence, or unauthorized data exfiltration were detected.
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of @aiagenta2z/onekey-gateway (NPM) and ai-agent-marketplace (PyPI). These are official vendor resources associated with the author 'aiagenta2z' and are necessary for the skill's functionality.
  • [COMMAND_EXECUTION]: The firecrawl_browser_execute tool provides an interface for running Bash, Python, or Node.js scripts within a remote browser session. This is an intended capability of the integrated Firecrawl service for web automation and does not represent a local vulnerability.
  • [CREDENTIALS_UNSAFE]: The skill follows best practices by using the DEEPNLP_ONEKEY_ROUTER_ACCESS environment variable for API authentication. A fallback demo key BETA_TEST_KEY_MARCH_2026 is provided for testing, which is standard for development-oriented skills.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 02:40 PM
Security Audit — agent-trust-hub — firecrawl-mcp