| Version | Supported |
|---|---|
| 1.x | Yes |
If you discover a security vulnerability in this project, please report it responsibly.
Do not open a public issue.
Instead, please use GitHub's private vulnerability reporting to submit your report.
You can expect:
- An acknowledgment within 48 hours
- A follow-up with an assessment within 7 days
- A fix timeline based on severity
This project is an MCP server that acts as a read-only client to the public SIMAP API. It does not handle authentication, store credentials, or process sensitive user data.
Relevant concerns include:
- Dependency vulnerabilities
- Input validation issues
- Unexpected data exposure through tool outputs
- Keep the package updated to the latest version
- Run the server in a sandboxed environment when possible
- Review tool outputs before acting on them
- Keep
SIMAP_MCP_DEBUGunset (or set to an empty value) in production. Debug mode logs the full request URL, including user-supplied search terms and filter values, to stderr. In most MCP hosts these stderr logs are captured and retained, so enabling debug in production effectively persists user queries. - The server does not require any secrets, tokens, or API keys. Do not configure any — if you see an example asking you to set one, it is likely a phishing attempt.
- stdout is reserved for the MCP JSON-RPC protocol; never redirect it into log files.
The SIMAP_MCP_DEBUG environment variable, when set to 1 or true, switches the HTTP client to verbose stderr logging:
- Full outbound URL (with all query parameters)
- Response status, byte size, and request duration
This is intended for local troubleshooting only. It is off by default precisely because the extra payload can contain user-intent data (search terms, CPV codes, canton filters) that should not leak into shared log infrastructure.