@one-source/mcp@5.18.0
Blockchain data, Deepstate and Standard Reserve markets, live chain queries, and API docs
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe credential-flow finding is a false alarm: the submitted OneSource key is validated against OneSource's own API. However, server creation enables runtime usage analytics by default, and the package documentation does not clearly disclose the collection; this is a warning-level telemetry concern.
Decision evidence
public snapshot- Creating the exported server enables analytics by default and sets its destination to https://1s-analytics.vercel.app; an environment variable can disable it.
- Tool calls send usage metadata such as tool and service names, timing, success, network, and input parameter names to the analytics endpoint.
- The analytics implementation posts buffered events to the endpoint's /api/events path.
- The flagged API key is submitted through the connect flow and validated against the configured OneSource API, defaulting to api.onesource.io.
- Analytics can be disabled with ONESOURCE_ANALYTICS=false.
- The analytics events shown contain usage metadata rather than tool argument values or API keys.
- The key validation sends the submitted key to the OneSource API, consistent with the package's connection flow.
Source & flagged code
4 flaggedSource appears to send environment or credential material to an external endpoint.
dist/connect-page.jsView on unpkg · L9A package entrypoint or install-time lifecycle script reaches a source file with blocking dangerous behavior.
dist/connect-page.jsView on unpkg · L9Findings
Affected versions and remediation
This report applies to @one-source/mcp@5.18.0.
See all versions of @one-source/mcp for other recorded verdicts.
- Review the evidence and your use of @one-source/mcp@5.18.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.