Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    a11y Review: Alt Text & Accessibility Fixer

    2

    Finds missing or meaningless alt text by actually looking at your images, then fixes it plus common HTML accessibility issues, only after you approve.

    $9

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    a11y Review: Alt Text & Accessibility Fixer

    a11y Review: Alt Text & Accessibility Fixer

    Example session with this skill installed

    Audit this site for accessibility issues. Check the images and the HTML structure.

    • Read your context and instructions
    • Compiled the a11y review
    • Generated the UI component

    Summary

    Scanned 3 files (index.html, Nav.astro, Newsletter.tsx) and 6 <img> tags. Nothing has been changed yet.

    • 4 images need alt text (3 verified by opening the image file, 1 inferred)
    • 2 images are already correct
    • 7 structural issues

    Image alt text (6 items)

    #File:LineIssueCurrentProposed altConfidence
    1index.html:19Missing(none)"Steaming latte in a white cup and saucer on a wooden table"Verified (image read)
    2index.html:28Placeholderalt="IMG_2031.jpg""Kraft paper bag of coffee beans with a cream label showing three beans"Verified (image read)
    3index.html:34Placeholderalt="image""Morning Cup Coffee storefront with a red-and-white striped awning"Verified (image read)
    4index.html:35Missing(none)"Spring campaign banner"Low confidence — remote URL, inferred from filename and context only
    5index.html:24Emptyalt=""alt="" (decorative wave divider)Verified — no change needed
    6index.html:41Emptyalt=""alt="" (icon beside the visible text "Instagram")Verified — no change needed

    Structural accessibility (7 items)

    #File:LineIssueCurrentFix
    7index.html:10<nav> has no aria-label; links not in a list<nav class="flex gap-4"> + bare <a> tags<nav aria-label="Main navigation" class="flex gap-4"> with the links wrapped in <ul><li>
    8Nav.astro:8<nav> has no aria-label; links not in a list<nav class="flex gap-4"> + mapped <a> tagsSame fix as #7, with <li> inside the .map()
    9index.html:20Second <h1> on the page<h1>Small-batch coffee, roasted every morning</h1><h2>…</h2> (the site title at line 9 is the page's <h1>)
    10index.html:21Link used as a button<a href="#" class="btn" onclick="openOrderDialog()"><button type="button" class="btn" onclick="openOrderDialog()">
    11index.html:27Heading level skips (h1 → h3)<h3>Our beans</h3><h2>Our beans</h2>
    12Newsletter.tsx:4<input> has no label (placeholder only)<input type="email" placeholder="you@example.com" /><label htmlFor="email">Email</label> + <input id="email" type="email" placeholder="you@example.com" />
    13Newsletter.tsx:5<button> has no type<button onClick={…}>Subscribe</button><button type="button" onClick={…}>Subscribe</button>

    → Type "yes" to apply all, or list the numbers you want (e.g. 1 2 3 7). Item 4 is low confidence — say the number followed by your own wording to override (e.g. 4 Spring collection: up to 30% off).

    Next steps

    • Items 5 and 6 are already correct as they are; no action needed.
    • Replace the placeholder text in item 4 with the banner's real message once you have the image.

    index.html

    TSX · React component

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Most accessibility linters can tell you an alt attribute is missing. They can't tell you what it should say, because that takes looking at the image. This skill does the looking.

    It scans your project, resolves every img or Image tag to the real file on disk, opens the image, and writes a specific, context-aware alt text: "Kraft paper bag of coffee beans with a cream label", not "image" or "IMG_2031.jpg". Purely decorative images get alt="" instead of a forced description, which is the correct answer for screen readers.

    What it checks

    • Images: missing alt, empty alt, and placeholder alt (filenames, "image", "photo")
    • Navigation: nav elements without an aria-label, link lists not wrapped in ul/li
    • Buttons and links: buttons without a type attribute, links with href="#" used as buttons
    • Forms: inputs with no associated label, placeholder used as the only label
    • Headings: multiple h1 tags, skipped heading levels

    How it works

    1. A bundled Node.js scanner finds every image tag in one deterministic pass (no dependencies).
    2. The agent opens each image file and proposes alt text, with a confidence level per item. Remote or unresolvable images are flagged as low confidence rather than guessed silently.
    3. You get numbered tables of proposed fixes. Nothing is edited until you say which ones to apply ("yes", or "1 2 5").
    4. Approved fixes are applied and summarized in a result table.

    Works with HTML, Astro, React/JSX/TSX, Vue, and Svelte, with or without Tailwind. Large projects are grouped by section so the review stays readable.

    What it is not

    It is a focused checklist for the issues above, not a full WCAG audit. It doesn't test color contrast or keyboard behavior at runtime, so keep using axe / Lighthouse for those. Use it to fix the alt-text and markup problems that tools flag but can't fix for you.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Frontend & Web Apps.

    Genuinely unresolvable remote images or dynamic source expressions result in lower-confidence inferences from filenames and context rather than visual analysis.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Software
    Node.js

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    File Scopes

    **/*.html
    **/*.astro
    **/*.jsx
    **/*.tsx
    **/*.vue
    **/*.svelte

    Terminal: runs the bundled read-only scanner (node scripts/scan-images.js) to list image tags. Read Files: reads project source files and the image files they reference so alt text can describe the actual image. Write Files: edits source files only after you approve specific numbered items; nothing is written without confirmation. No network access and no environment variables.

    Listed1 month ago
    Updated1 month ago

    What's inside

    Frequently Asked Questions

    Linters such as eslint-plugin-jsx-a11y can tell you an alt attribute is missing or empty, but not what it should say. This skill opens each image file and proposes alt text based on what is actually in the picture. It also works on HTML, Astro, Vue, and Svelte, not just JSX. It complements a linter rather than replacing it, so keep running yours.

    HTML, Astro, React (JSX/TSX), Vue, and Svelte, with or without Tailwind. It needs an agent that can read and edit files and run a Node.js script, and, for image-based alt text, one that can view image files, such as Claude Code. No other packages are required.

    The agent needs to read your project's source files and the image files they reference, and to run the bundled Node.js scanner. It edits files only after you approve specific numbered items. It needs no network access and no environment variables.

    It looks at the image itself and the surrounding markup, such as a nearby heading, caption, or link text. Purely decorative images, like a background pattern or an icon beside redundant visible text, get alt="" instead of a forced description. Every proposal is listed for your approval first, so you can override any of them.

    You get the skill itself (SKILL.md), the bundled Node.js image scanner (scripts/scan-images.js), and an MIT license file. There are no other files or dependencies.

    No. It always shows a numbered list first and applies only the items you select.

    It can't open them, so it infers from the filename and surrounding markup and marks those suggestions low confidence.

    No. It covers image alt text and the structural checks listed above. Pair it with axe or Lighthouse for contrast and runtime checks.

    New in Frontend & Web Apps