Google Modern Web Guidance 1.0: free skill pack for AI coding agents

Published November 11, 2026

A MacBook with lines of code on its screen on a busy desk

Google shipped version 1.0 of Modern Web Guidance, a free skill set that web teams can drop into their AI coding agents to steer output toward current browser capabilities rather than outdated patterns. The project is MIT-territory free, installable in one command, and now backed by Chrome, Microsoft Edge, and Mozilla.

Table of Contents

What It Is and How to Install It

Modern Web Guidance packages browser-current guidelines as agent-readable rules. When a coding agent generates code, it checks the guidance before writing and produces output evaluated against real browser metrics: Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and accessibility checks.

Install it by running the following in any project directory:

npx modern-web-guidance@latest install

It’s also available through the plugin marketplaces of most popular coding agents, according to the announcement from Philip Walton, a senior staff developer relations engineer at Google. Preview users can update with:

npx modern-web-guidance@latest update

What the Evals Actually Show

Google runs a public eval dashboard where coding agents from Anthropic, OpenAI, and Google run identical tasks nightly, with and without the guidance loaded. A real browser then checks the DOM, styles, and source code of each finished app.

Based on nightly runs completed November 8, guided runs passed 88% of automated checks. Unguided runs passed 56%. Each agent ran between 199 and 215 tasks. Guided runs were slower and burned more tokens, according to the dashboard data. The task prompts do not mention the guidance, so each agent has to work out on its own when a guide applies.

The dashboard includes a limited accessibility review covering keyboard support, focus management, and form labels. It is not a full WCAG audit, per the dashboard’s own disclosure.

️ What’s New in v1.0

Codebase audit: scans HTML, CSS, JavaScript, and package.json to flag areas where safe updates are possible, then walks the agent through the changes. This first iteration uses static analysis on a limited set of code patterns.
Expanded guides: the library now targets hundreds of verified AI knowledge gaps across its guides.
Multi-browser governance: a governance document uploaded November 9 defines six owner seats: two for Chrome, one each for Microsoft Edge and Mozilla, and two for independent web platform experts. One independent seat is currently open. Apple holds no seat.
Domain review process: each guide is now reviewed and approved by a domain tech lead. User-facing feature guides include a dedicated accessibility review by experts who use assistive technologies daily.

Early Adopter Results

Japan-based creator platform note inc. ran two coding agents through six UI tasks across 180 test runs on clone versions of its production codebase. Both agents improved on 19 browser checks in three stages: unguided, guided without naming the guidance in prompts, then guided with the guidance named explicitly.

Agent One: 60.8% → 76.5% → 94.0%
Agent Two: 75.8% → 86.2% → 90.0%

MetaMask runs the guidance as an automatic reviewer on every pull request to its open-source browser extension. Through October 21 (the first five weeks), the bot flagged 102 issues across 57 pull requests. Engineers fully resolved 43, partially resolved 3, declined or marked as false positives 16, and left 20 unaddressed. The outcome of the remaining 20 is unknown.

The Catch

Accessibility consultant Adrian Roselli evaluated the May preview using Google’s AI coding tool Antigravity and found the output did not comply with WCAG, with animations that were non-functional in Firefox. He also found no assurance that agents would reliably follow the guidelines. That evaluation was of the preview, not v1.0, but it is worth factoring in before relying on the guidance as a compliance backstop rather than a quality nudge.

Google suggests pairing Modern Web Guidance with Chrome DevTools for agents so the agent can run and test its own output directly in the browser.

Cagri Sarigoz( Founder )

I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.

With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.

At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.

And more than all else, I’m a father to a (always) little girl and a husband.

Contact me at [email protected] for assistance.