Prevent DOM-based XSS attacks with Trusted Types

Trusted Types is a security feature that helps prevent DOM-based Cross-Site Scripting (DOM XSS) by requiring that data being passed into “dangerous” browser APIs (known as sinks) is first converted into a specific type, such as TrustedHTML.

By default, many DOM APIs (like innerHTML or document.write) accept raw strings. If a string contains malicious script from an untrusted source, it can be executed. Trusted Types changes this behavior: once enabled via CSP, these sinks will only accept TrustedHTML objects created by authorized policies.

Key Implementation Details

To use Trusted Types, follow four main steps:

1. Define a security policy that specifies how to sanitize or transform strings.
2. Use that policy to create Trusted Type objects from raw strings.
3. Use the Trusted Type objects when writing to DOM sinks.
4. Enforce Trusted Types using a Content Security Policy (CSP).

1. Creating a Policy

A policy is a set of rules for sanitizing content. You create it using trustedTypes.createPolicy(), typically delegating to a sanitizer library such as DOMPurify or your application’s HTML escaping rules:

import DOMPurify from 'dompurify';

const myPolicy = trustedTypes.createPolicy('sanitize-html', {
  createHTML: (input) =>
    // Return a sanitized string; createPolicy() wraps it in a TrustedHTML object.
    // Example using DOMPurify:
    DOMPurify.sanitize(input, { RETURN_TRUSTED_TYPE: false }),
});

// Example using HTML escaping (for environments without third-party sanitizers):
const escapePolicy = trustedTypes.createPolicy('escape-html', {
  createHTML: (input) =>
    String(input)
      .replaceAll('&', '&')
      .replaceAll('<', '&lt;')
      .replaceAll('>', '&gt;')
      .replaceAll('"', '&quot;')
      .replaceAll("'", '&#39;'),
});

[!WARNING] Character-level HTML escaping (escapePolicy) converts all markup into plain text and is not a context-aware HTML sanitizer. If you only need plain text, prefer element.textContent (which is not a Trusted Types sink). Never interpolate escapePolicy.createHTML() output into another HTML template string—such as an unquoted attribute (<img src=${escaped}>) or a URL attribute (<a href="${escaped}">, where javascript: URLs are not neutralized by entity escaping). Use a full HTML sanitizer (DOMPurify or element.setHTML(), covered in sanitize-untrusted-html) whenever accepting HTML markup or attributes.

2. Using the Policy

Instead of passing a string directly to a sink, you call the policy’s creation method.

const untrustedInput = '<p>Hello, <strong>world</strong>!</p><img src="x" onerror="alert(1)">';

// In an enforced environment, this would throw a TypeError:
// element.innerHTML = untrustedInput;

// This returns a TrustedHTML object with unsafe elements and attributes removed:
const cleanHTML = myPolicy.createHTML(untrustedInput);
// Result: "<p>Hello, <strong>world</strong>!</p><img src=\"x\">"

3. Transitioning sinks to use Trusted Types

Update each instance where you write to a sink to first convert it to a Trusted Type object. To help locate where code needs to be updated, you can define a temporary default policy that logs when a sink is written to with a string rather than a TrustedType.

trustedTypes.createPolicy("default", {
  createHTML(value) {
    console.warn("String passed instead of Trusted Type.");
    return DOMPurify.sanitize(value, { RETURN_TRUSTED_TYPE: false });
  },
});

Common sinks that support Trusted Types include:

• HTML: element.innerHTML, element.outerHTML, document.write(), element.setHTMLUnsafe(), Document.parseHTMLUnsafe()
• Script: <script> text content, eval()
• ScriptURL: <script src>, Worker(), SharedWorker()

// Assign the TrustedHTML object.
element.innerHTML = cleanHTML;

Note that the “safe” HTML sanitization methods (element.setHTML() and Document.parseHTML(), covered in sanitize-untrusted-html), do not require a Trusted Types policy, as they always sanitize potential XSS attacks natively.

4. Enforcing Trusted Types with CSP

Trusted Types should be rolled out incrementally to avoid breaking your application. While full enforcement provides the strongest protection, it carries a risk of site breakage if every sink write hasn’t been migrated.

Prerequisites

• Framework cooperation: If your application’s framework (or any third-party library that writes to DOM sinks) does not produce TrustedHTML or TrustedScript values, you cannot enforce Trusted Types without breaking that code. Audit framework support before starting the rollout.
• Sink refactoring: Ensure you have identified and refactored dangerous sinks to use your Trusted Types policies.

Rollout Strategy

1. Report-Only: First, roll out using the Content-Security-Policy-Report-Only header. This will log violations to your reporting endpoint without actually blocking the writes.

   Content-Security-Policy-Report-Only: require-trusted-types-for 'script';

2. Enforcement: Once you have integrated your policies and violation reports drop to zero, move to full enforcement.

   Content-Security-Policy: require-trusted-types-for 'script';

You can also restrict which policies are allowed to be created (list your policy names such as sanitize-html or escape-html, include dompurify if your policy delegates to DOMPurify.sanitize()—which internally creates a dompurify policy—and include default if you register a default policy during migration):

Content-Security-Policy: require-trusted-types-for 'script'; trusted-types sanitize-html escape-html dompurify;

With this CSP in place, any writes to a sink that don’t use Trusted Types will throw an error. In addition, attempting to create a Trusted Type policy with a name not listed in trusted-types will throw an error.

You can also set this CSP as a <meta> tag in your document’s <head>, although this is potentially less secure.

<!-- First child of `<head>`, as earlier tags will not enforce this CSP. -->
<meta
  http-equiv="Content-Security-Policy"
  content="require-trusted-types-for 'script'; trusted-types sanitize-html escape-html dompurify;"
/>

Fallback strategies

Baseline: trusted-types

For browsers that do not yet support Trusted Types, you can use a minimal “tinyfill” to ensure your code doesn’t break while still providing sanitization.

if (!window.trustedTypes) {
  window.trustedTypes = {
    createPolicy: (name, rules) => rules,
  };
}

This allows you to write your code using the Trusted Types pattern today. Supporting browsers will enforce the types, while older browsers will still execute your sanitization logic via the mock policy.

If a browser doesn’t support Trusted Types, it will simply treat the TrustedHTML objects as strings (if using the tinyfill) or won’t have the enforcement check at the sink level.

Compatibility notes: Badges like Baseline: … mark web features referenced by this guide. Baseline status describes availability across the core browser set — it does not mean every browser or older version supports the feature. Feature data as published upstream, synced 2026-11-12.

Live demo

Open the demo standalone · Demo source on GitHub

---

Source: GoogleChrome/modern-web-guidance-src (Apache-2.0), upstream commit 179dbc6, synced 2026-11-12. Local changes: macros expanded to links/badges, first heading de-duplicated, demo and this footer added. This is an independent, unofficial site — not a Google publication.