Why this verdict
- benign: 331 checks satisfied, no findings for mcp-server:.
Security findings 0
No security findings
No suspicious or malicious behavior was found; 1 candidate finding was refuted on review.
Execution chain 24
What it calls, runs, installs, reads and writes, with the line that does it.
GitHub repositories 4
willianpinho/mcp-gateway-scan
willianpinho/mcp-gateway-scan
willianpinho/mcp-gateway-scan
willianpinho/mcp-gateway-scan
Packages 1
npm:mcp-gateway-scan@0.2.0
URLs 7
https://github.com/willianpinho/mcp-gateway-scan
https://github.com/willianpinho/mcp-gateway-scan.git
https://github.com/willianpinho/mcp-gateway-scan/issues
https://github.com/willianpinho/mcp-gateway-scan#readme
https://provenwright.com/audit
https://registry.npmjs.org
https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json
Refuted findings 1
Candidates the review stage examined and rejected, kept for transparency.
Reconnaissance
lowRefutedlow confidenceThe scan_gateway MCP tool resolves an agent-supplied path with no workspace restriction, and the walker reads .env files and other config files under it, so an agent can enumerate arbitrary directories and get matched-line excerpts back in the tool result (read-only, no network sink).
const target = resolve(inputPath);
Why it was refuted
The tool is a directory scanner whose purpose is to take a path and scan it, so reading config and .env files under that path is its core function. The output goes only to the calling agent, with no network sink or exfiltration chain.