Lightweight TypeScript library for validating URLs to prevent Server-Side Request Forgery (SSRF) attacks. It blocks requests to private IP ranges, localhost, link-local addresses, and other special-purpose networks using Microsoft's maintained @microsoft/antissrf block lists.

About

node-safeurl provides the following exports:

• fetch(input, init?) — SSRF-safe drop-in replacement for the global fetch. Validates the URL (and every redirect hop) before making the request, and returns a TypedResponse. Pass { safeUrl: false } to opt out of validation entirely, or { safeUrlAllow: [...] } to allow specific private-network origins while keeping all other SSRF checks active.
• TypedResponse — Subclass of Response that adds a .typed(schema) method for runtime-validated JSON parsing via TypeBox.
• FetchInit — TypeScript interface extending RequestInit with the optional safeUrl boolean and safeUrlAllow string-array fields.
• isSafeUrl(href, opts?) — Async function that validates a URL is safe to fetch. Checks protocol, hostname, IP literals, and performs DNS resolution to guard against DNS rebinding attacks. Pass { allow: [...] } to exempt specific origins or hostnames from all checks.
• SafeUrlOptions — TypeScript interface for the options accepted by isSafeUrl.
• isPrivateIPv4(address) — Synchronous check for private/special-purpose IPv4 addresses.
• isPrivateIPv6(address) — Synchronous check for private/special-purpose IPv6 addresses.

Installation

NPM

npm install @tak-ps/node-safeurl

Usage

import fetch, { isSafeUrl, isPrivateIPv4, isPrivateIPv6 } from '@tak-ps/node-safeurl';
import { Type } from '@sinclair/typebox';

// SSRF-safe fetch — validates the URL and every redirect hop automatically
const res = await fetch('https://example.com/api/data');
const data = await res.typed(Type.Object({ id: Type.Number() }));

// Opt out of SSRF validation entirely for trusted internal calls
const internal = await fetch('http://localhost:3000/health', { safeUrl: false });

// Allow a specific private-network origin while keeping all other SSRF checks active
const api = await fetch('http://10.0.0.5:8080/api/status', {
    safeUrlAllow: ['http://10.0.0.5:8080'],
});

// Validate a URL manually before fetching
const result = await isSafeUrl('https://example.com/api');
if (result.safe) {
    // Safe to fetch
    const response = await fetch(result.url);
} else {
    console.error('Blocked:', result.reason);
}

// Validate a URL while allowing a known private endpoint
const result2 = await isSafeUrl('http://192.168.1.10/health', {
    allow: ['192.168.1.10'],
});

// Check individual IPs
isPrivateIPv4('192.168.1.1');  // true
isPrivateIPv4('8.8.8.8');      // false
isPrivateIPv6('::1');          // true
isPrivateIPv6('2606:4700:4700::1111'); // false

API

fetch(input, init?): Promise<TypedResponse>

SSRF-safe drop-in replacement for the global fetch. Validates the initial URL and every redirect destination against isSafeUrl before the request is made. Throws an Err(403) if a URL is deemed unsafe.

init accepts all standard RequestInit options plus:

• safeUrl (boolean, default true) — set to false to skip all SSRF validation (e.g. for fully trusted internal endpoints).
• safeUrlAllow (string[], default []) — list of origins (e.g. "http://10.0.0.5:8080") or bare hostnames (e.g. "10.0.0.5") that bypass SSRF checks while all other URLs remain validated. Use this instead of safeUrl: false when only specific private-network endpoints need to be exempted.

Custom dispatcher options are rejected when safeUrl is true because they can bypass SSRF protection.

TypedResponse

Subclass of Response returned by fetch. Adds:

.typed<T>(schema: TSchema): Promise<Static<T>>

Parses the response body as JSON and validates it against a TypeBox schema. Throws Err(500) if validation fails.

const res = await fetch('https://api.example.com/user/1');
const user = await res.typed(Type.Object({
    id: Type.Number(),
    name: Type.String(),
}));

FetchInit

TypeScript interface extending RequestInit with additional fields:

interface FetchInit extends RequestInit {
    safeUrl?: boolean;      // default: true — set false to skip all SSRF checks
    safeUrlAllow?: string[]; // origins or hostnames exempt from SSRF checks
}

isSafeUrl(href: string, opts?: SafeUrlOptions): Promise<{ safe: boolean; url?: URL; reason?: string }>

Validates that a URL is safe to fetch from a server context. Returns an object with:

• safe — true if the URL is safe, false if it should be blocked
• url — The parsed URL object (when the URL could be parsed)
• reason — A human-readable string explaining why the URL was blocked

opts accepts:

• allow (string[]) — origins (e.g. "http://10.0.0.5:8080") or bare hostnames that are unconditionally considered safe, bypassing all checks below.

Checks performed (when not matched by allow):

1. URL must be parseable
2. Protocol must be http: or https:
3. Hostname must not be localhost or 0.0.0.0
4. IP literal hostnames must not be in private/special-purpose ranges
5. DNS resolution results must not map to private/special-purpose IPs

SafeUrlOptions

interface SafeUrlOptions {
    allow?: string[]; // origins or hostnames exempt from all SSRF checks
}

isPrivateIPv4(hostname: string): boolean

Returns true if the given string is a valid IPv4 address in a private or special-purpose range.

isPrivateIPv6(address: string): boolean

Returns true if the given string is a valid IPv6 address in a private or special-purpose range.

Blocked Ranges

The following ranges are blocked (via @microsoft/antissrf):

• IPv4: Loopback (127.0.0.0/8), RFC 1918 (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16), Link-local (169.254.0.0/16), CGNAT (100.64.0.0/10), and more
• IPv6: Loopback (::1/128), ULA (fc00::/7), Link-local (fe80::/10), IPv4-mapped (::ffff:0:0/96), and more

License

MIT