Security Policy

This document describes the security policy for the SIMAP MCP Server, including supported versions, vulnerability reporting procedures, security scope, and best practices. The SIMAP MCP Server is designed as a read-only client to the public SIMAP.ch API with minimal security surface area.

For information about community conduct and behavior standards, see Code of Conduct. For details on the contribution process and PR requirements, see How to Contribute.

Sources: SECURITY.md1-36 CODE_OF_CONDUCT.md1-83

---

Supported Versions

The project maintains security support for the following version ranges:

Version	Supported	Notes
1.x	✅ Yes	Current major version, receives security updates SECURITY.md5-7
< 1.0	❌ No	Pre-release versions, upgrade to 1.x

Security patches are released as patch version increments (e.g., 1.2.2 → 1.2.3) and automatically published through the CI/CD pipeline. Users should monitor GitHub releases or npm updates to stay current with security fixes.

Sources: SECURITY.md3-7

---

Vulnerability Reporting Process

Reporting Channel

Security vulnerabilities must be reported through private channels only. Do not open public GitHub issues for security concerns SECURITY.md13

Primary reporting method: Use GitHub's private vulnerability reporting to submit confidential security reports SECURITY.md15

Alternative contact: For issues requiring direct communication or reporting Code of Conduct violations, contact project maintainers at info@digilac.ch CODE_OF_CONDUCT.md49

Vulnerability Reporting Workflow

Title: "Vulnerability Management Lifecycle"

Expected Response Times

Stage	Timeline	Description
Acknowledgment	48 hours	Confirming receipt of vulnerability report SECURITY.md18
Initial Assessment	7 days	Determining validity and severity SECURITY.md19
Fix Timeline	Variable	Based on severity and impact SECURITY.md20

Sources: SECURITY.md9-21

---

Security Scope and Threat Model

System Security Boundaries

The SIMAP MCP Server operates within clearly defined security boundaries, acting as a bridge between an MCP Client and the public SIMAP API.

Title: "System Security Boundaries and Code Entities"

In-Scope Security Concerns

The following security areas are within the project's scope SECURITY.md26-30:

1. Dependency Vulnerabilities: Security flaws in third-party libraries like @modelcontextprotocol/sdk or zod.
2. Input Validation Issues: Ensuring that user-provided parameters to tools do not cause unexpected server behavior.
3. Unexpected Data Exposure: Ensuring tool outputs do not leak sensitive information or excessive internal data.

Out-of-Scope Security Concerns

Area	Rationale
Authentication/Authorization	Server is a read-only client to a public API; it does not handle auth SECURITY.md24-38
Credential Storage	The server does not store or manage any user credentials SECURITY.md24-40
User Data Privacy	The server does not process or store sensitive user data SECURITY.md24
Write Operations	The server only performs read operations against public endpoints SECURITY.md24-38

Sources: SECURITY.md22-40

---

Production Deployment Guidance

Debug Mode and Data Leakage

The SIMAP_MCP_DEBUG environment variable controls verbose logging to stderr. This is intended for local troubleshooting only SECURITY.md45-50

• Behavior: When set to 1 or true, it switches the HTTP client to verbose logging, including full outbound URLs with all query parameters (search terms, filters, CPV codes) and response metadata SECURITY.md45-50
• Risk: Most MCP hosts capture stderr logs and retain them. In production, this can effectively persist user-intent data in shared log infrastructure SECURITY.md39-50
• Requirement: Keep SIMAP_MCP_DEBUG unset or empty in production environments .env.example5-8 SECURITY.md39

Secrets and Credentials

The server does not require any secrets, tokens, or API keys SECURITY.md40

• Phishing Alert: If any documentation or example requests you to set an API key or secret for this server, it is likely a phishing attempt SECURITY.md40
• Protocol Integrity: stdout is strictly reserved for the MCP JSON-RPC protocol. Never redirect stdout into log files as it will break the communication with the MCP client SECURITY.md41

Sources: SECURITY.md37-51 .env.example1-9

---

Security Architecture and Mitigations

Input Validation Layer

All tool inputs pass through strict Zod schema validation before processing. This prevents injection attacks and ensures the API client receives only expected data types.

Title: "Input Validation Pipeline"

HTTP Client Security

The SimapClient class implements security best practices to protect the server and the upstream API:

Title: "SimapClient Security Controls"

Sources: SECURITY.md22-30 CODE_OF_CONDUCT.md45-53

---

Best Practices for Users

Installation and Updates

1. Keep Updated: Always use the latest version to ensure you have the most recent security patches SECURITY.md33
2. Sandboxed Execution: Run the server in a sandboxed environment (like a container or restricted user account) to limit the impact of any potential vulnerability SECURITY.md34
3. Output Review: Review tool outputs before acting on them, especially when the output is used to inform further automated actions SECURITY.md35

Sources: SECURITY.md31-36