Chrome's Modern Web Guidance hits 1.0, adds Firefox and Edge

Chrome has released the stable 1.0 version of Modern Web Guidance, a set of skills that folds web platform expertise, best practices and browser compatibility data directly into AI coding agents. The release matters because it targets the exact spots where even strong frontier models still ship web code that is slower, less accessible or less secure than an expert would write — and because Firefox and Edge are now joining Chrome to contribute to the project. It is available now as an installable package for existing agent workflows.

What is Modern Web Guidance?

According to Chrome for Developers, Modern Web Guidance is a set of skills that embed web platform expertise, best practices and browser compatibility data directly into coding agents. The Chrome team first introduced an early preview at Google I/O in May. The problem it targets, as the announcement frames it, is that today's frontier coding models are capable and can usually produce something that works, but a web development expert reviewing that AI-generated code can almost always find ways to make it faster, more accessible or more secure. Modern Web Guidance is meant to hand the agent that expertise up front, and the team says it works regardless of which models or agent harnesses you use. Although it is built primarily as a tool for agents, the team notes that many developers have found the guides useful on their own as a learning resource, either by reading them or by using the CLI to search use cases and discover new techniques.

What does version 1.0 add?

The stable release adds four things, according to the post. It introduces a modernization workflow that scans your existing codebase to help an agent find opportunities to safely modernize. It expands the guide database to target hundreds of verified AI knowledge gaps with expert-vetted, cross-browser solutions. It deepens human expert review for every guide, including sign-off from content-area tech leads and a dedicated accessibility review on user-facing guides. And it adds full evaluation coverage across all guides, with a public eval dashboard showing measurable improvements compared with agents that do not use Modern Web Guidance. The release also marks a bigger shift: Modern Web Guidance is becoming a cross-browser effort, with Firefox and Edge joining Chrome to support the project, improve cross-browser compatibility and contribute. The work builds on the shared Baseline effort.

Where do coding agents still get web development wrong?

Chrome's evals, as described in the post, point to recurring gaps. In accessibility, models struggle with keyboard focus and inert management, reduced-motion support and accessible form validation. In user experience, they miss swipe and touch gestures, dialog and popover transitions, and light/dark theme implementations. On performance, the shortfalls include main-thread task scheduling, offscreen rendering isolation and layout shift prevention. On security, they cover safe HTML sanitization, strict Content Security Policies and passkey credential management. Accessibility is singled out as especially important, because bad ARIA can leave a page less accessible than no ARIA at all; the user-facing guides were reviewed and tested with real assistive technologies by experts who use them daily. The guidance also nudges agents toward the best web feature for a job — native HTML and CSS for popovers, dialogs, gestures and page transitions instead of hand-written JavaScript, plus passkeys, WebAuthn and autofill for login and checkout.

How does the modernization workflow work?

Most web development happens in existing codebases that have gathered years of outdated workarounds, unnecessary dependencies and no-longer-needed polyfills, the post notes. The new workflow analyzes a project's HTML, CSS, JavaScript and package.json to find opportunities to safely upgrade, then guides the agent through the changes. You start it with a command such as /modern-web-guidance Modernize the code in this directory. In this first release, the workflow uses static analysis to check a limited set of patterns; the team chose a narrow scope for predictable results and plans to expand over time, including agentic audits where the agent reasons about the code directly instead of only matching known patterns. Because the guides carry embedded Baseline and browser-compatibility data and are maintained by browser engineers, standards experts and accessibility specialists, the recommendations evolve as the platform does.

How does it pair with Chrome DevTools for agents?

Chrome recommends using Modern Web Guidance alongside Chrome DevTools for agents. The division of labor is straightforward: Modern Web Guidance gives the agent the up-front expertise to write code that follows modern best practices, while Chrome DevTools for agents gives it runtime visibility to verify, debug and optimize that code across a range of devices, user preferences and network conditions. The post's examples include running a performance trace on a simulated low-end CPU to keep main-thread tasks under 50 milliseconds so they do not contribute to poor INP, recording animations to confirm they work and honor prefers-reduced-motion, and using memory tools to ensure cached single-page-app views do not leak memory over a long session. Together, the team says, the two tools help an agent arrive at a better solution in fewer turns.

How could you explore this with Grafy?

Modern Web Guidance is a toolkit for coding agents, and the closest fit in Grafy is Grafy Code, the AI coding studio that writes and runs your code. If you build web projects there, the practices this release promotes — accessible markup, native CSS for interactions, and performance and security hygiene — are useful for judging what an agent produces. Keep in mind that Modern Web Guidance is a separate, installable package; Grafy Code does not ship it built in. To study the ideas more closely, you could drop the announcement or the guides into Grafy Reader to question the details, or compare approaches across several models side by side in Grafy Chat.

Frequently asked questions

Does Modern Web Guidance work with any coding model or agent?

Yes. According to Chrome for Developers, the guidance is designed to work regardless of which models or agent harnesses you use. It embeds web platform expertise, best practices and Baseline browser-compatibility data as skills, so the recommendations apply whether you are using a frontier model or a different setup, as long as you are building for the web.

What does version 1.0 add over the earlier preview?

The stable release adds a modernization workflow that scans existing code, an expanded guide database covering hundreds of verified knowledge gaps, deeper human expert review with accessibility sign-off, and full evaluation coverage with a public eval dashboard. It also marks the project becoming a cross-browser effort, with Firefox and Edge joining Chrome to contribute.

How do you install or update Modern Web Guidance?

According to the announcement, you install it by running npx modern-web-guidance@latest install from your project directory. If you already have an earlier version, you update to 1.0 with npx modern-web-guidance@latest update. Chrome also points to the Modern Web Guidance documentation and an overview video for more detail.

Source: Chrome for Developers

---

Source: Hacker News — front page — Modern Web Guidance 1.0: The web platform expertise your coding agent is missing