Use when a feature needs manual screen-reader testing before sign-off: step-by-step scripts for NVDA, JAWS, VoiceOver or TalkBack with keystrokes and expected announcements, WCAG 2.2 checklists per page type, widget keystroke matrices and a guided session with a signed report.
94%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Low
Low-risk findings worth noting
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
The required workflow uses a user-supplied rendered URL during guided testing sessions to check screen-reader accessibility, allowing an outsider to control content fetched at runtime.
user-supplied rendered URL (web page content)
content-type · 5 sites
The plugin requires a user-supplied rendered URL as a mandatory input and instructs the agent to open it in a browser and interact with its DOM content (reading headings, form fields, announcements), exposing the agent to arbitrary third-party web content that could contain indirect prompt injection.
SKILL.md
171
Required inputs: the component's rendered URL, its
SKILL.md
173
Without a rendered URL the pre-conditions cannot be defined - stop
SKILL.md
180
Header: component name, archetype, URL, date, tester name (blank field),
SKILL.md
200
Pre-conditions block: SR running, browser open to the URL, virtual
SKILL.md
222
**Component:** <name> **Archetype:** <archetype> **URL:** <url>