# Security

Hardening guidance for FastRMCP deployments. Pair with [Middleware Cookbook](Middleware-Cookbook.md) and [Transports Guide](Transports-Guide.md).

## Auth Patterns
- **Bearer token**: `AuthMiddleware::bearer_token("secret")` for quick protection; extend with custom middleware to validate per-request metadata.
- **Custom middleware**: Inspect `request.params.metadata` (or `_meta`) for headers/tokens; return `MCPError::InvalidRequest("unauthorized")` on failure.
- **Client identity**: Store authenticated user in `Context` metadata (`context.set_metadata("user", json!(user_id)).await`) for downstream tools.
- **Transport auth**: For web transports, add Axum middleware before the MCP router to verify headers/cookies.

## Rate Limiting
- Use `RateLimitMiddleware::new(limit, window)` (simplified) or custom middleware with shared counters keyed by `connectionId`/client ID.
- Apply stricter limits on expensive tools/resources; consider per-URI token buckets.

## CORS & Exposure
- Restrict origins on web transports; replace permissive `CorsLayer::new().allow_origin(Any)` with explicit origins.
- Expose only required transport(s) publicly; keep STDIO/internal transports behind firewalls or run on localhost.
- Prefer TLS for SSE/WS in production; see `docs/HTTP2.md` for rustls setup.

## Data Handling
- Validate input types explicitly; use `InvalidParams` for user errors and avoid leaking internals.
- Avoid returning sensitive metadata in tool results; strip in `after_response` middleware if needed.
- For resources backed by filesystem/network, sandbox paths and timeouts; return `ResourceError` with generic messages on failure.

## Connection IDs
- SSE: reject POSTs without `connectionId` when multiple connections exist; never echo other clients’ IDs.
- WebSocket: server-initiated requests default to current connection; set `metadata.connectionId` to target explicitly and avoid cross-tenant leaks.

## Operational Tips
- Rotate secrets without restart by storing them in shared `State` and reading inside middleware.
- Log auth failures with minimal detail; avoid logging credentials.
- Add health checks outside MCP if exposing HTTP servers; keep MCP methods private to trusted clients.

Related: [Observability](Observability.md) for logging failed auth/rate limits • [Troubleshooting](Troubleshooting.md) for common connection errors.

[Back to Home](Home.md)