Optimizing Claude Code Accessibility Prompts with axe-core Rule IDs

On November 4, 2026, the AI section of DEV Community published an article titled "Use axe-core violation lists to optimize accessibility prompts for Claude Code". It describes a prompting approach for improving accessibility fixes generated by Claude Code. According to the author, when a user sends accessibility repair instructions to Claude Code, vaguely worded prompts tend to make the agent blindly add far too many aria-label attributes. The best practice the article recommends is to feed the specific axe-core violation rule IDs and CSS selectors into the prompt as inputs, and to validate the results with Playwright tests. The article's body includes a set of background statistics from WebAIM: in February 2026, WebAIM scanned the home pages of a million websites, and 82.7% of them used ARIA, averaging more than 133 ARIA attributes per page. Pages using ARIA averaged 59.1 detected errors, while pages without ARIA averaged 42. The author notes that WebAIM is careful to point out that the ARIA-using pages were also more complex, so the data does not by itself establish ARIA as the cause of the errors. The article's byline material also lists related work by the same author, including a programming guide and the Hermes IDE project, a GitHub project aimed at developers who ship with Claude Code and other AI coding tools. As of the available sources, this is the only public report on the topic, and no further developments have been documented.