# Audit AI Features for Accessibility Compliance AI Agent Connect

> AI Feature Accessibility Scorer assesses the accessibility health of AI-driven software features. This MCP analyzes compliance against WCAG standards, checks alt text quality for visual assets, and validates keyboard navigation patterns. Your agent connects once through Vinkius and gets immediate access to detailed reports on technical failures, ensuring your product works for everyone. Use the `calculate_feature_score` tool for a high-level health assessment, `analyze_content_description` to check if AI-generated alt text is sufficient, `check_navigation_flow` to detect keyboard focus traps, and `get_compliance_gaps` to identify specific technical failures.

## Overview
- **Category:** accessibility
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_rV2BpIxzG7NdjZiOF2bymyAgpzUXW5625eSX7XqP/ai-agent-connect
- **Tags:** wcag, accessibility, ai-safety, compliance, ux-audit

## Description

Launching an AI feature is only half the battle. If it's unusable for people who rely on screen readers or keyboard navigation, you've got a compliance problem. This MCP gives you specialized diagnostic tools to audit your AI software features against WCAG standards. Instead of guessing, you get hard data. You can run a high-level health check, validate that your AI-generated descriptions for images are detailed enough, and confirm that every interaction point is reachable using only a keyboard. This MCP helps you find specific technical failures, like focus traps, before a user ever hits the bug report button. It’s about making sure your powerful AI features are genuinely usable by all.

## Tools

### calculate_feature_score
Provides the primary high-level accessibility health assessment for a specific AI feature

### check_navigation_flow
Validates if the AI interaction patterns are usable via keyboard

### get_compliance_gaps
Retrieves a detailed list of specific failures against the selected WCAG standard

### analyze_content_description
Evaluates the quality of text descriptions generated by the AI for visual assets

## Prompt Examples

**Prompt:** 
```
What is the accessibility score for feature 'chat-ui-01' at WCAG AA level?
```

**Response:** 
```
The accessibility score for 'chat-ui-01' at WCAG AA level is 85. There are 2 minor compliance gaps identified.
```

**Prompt:** 
```
Is the following description sufficient for an image of a sunset: 'A bright orange sun setting over the ocean'?
```

**Response:** 
```
The description quality score is 75. It provides basic clarity but lacks detail regarding the colors of the sky or the movement of the water.
```

**Prompt:** 
```
Check if the navigation flow for 'image-gen-tool' is keyboard accessible.
```

**Response:** 
```
The navigation flow is keyboard navigable. No focus traps were detected in the provided interaction steps.
```

## Capabilities

### Feature Health Scoring
Use this when you need a single, high-level score to gauge overall accessibility compliance for a new AI feature.

### Keyboard Flow Validation
Run this when you suspect users might get stuck or unable to navigate a feature using only the Tab key.

### Gap Identification
Call this when you need a granular, technical list of specific WCAG failures instead of just a score.

### Alt Text Quality Check
Run this when you want to confirm that the AI-generated descriptions for images are detailed enough for screen readers.

## Use Cases

### Pre-Launch Audit
Before releasing a new AI tool, run a full audit to get a feature score and check the navigation flow. This prevents a public failure.

### Image Generation Review
When your AI generates images, use the content description tool to ensure the alt text is rich and detailed enough.

### Compliance Reporting
Need to prove WCAG adherence? Use the gap identification tool to pull a detailed list of every technical failure.

### UX Iteration
After a round of user testing, run the navigation flow check to see if the new interaction patterns are still keyboard accessible.

## Benefits

- You get a specific score, not just a vague feeling, about your feature's accessibility health.
- It detects technical failures, like focus traps, that standard testing might miss.
- You confirm that AI-generated visual descriptions are detailed enough for screen reader users.
- The tool provides a clear list of compliance gaps, telling you exactly what needs fixing.

## How It Works

Connect your AI client to this MCP via Vinkius. You then invoke a specific tool, providing the feature or content you want audited. The MCP runs the WCAG diagnostic and returns a structured report detailing the score, gaps, or pass/fail status.

1. Connect your preferred AI client to the Vinkius catalog.
2. Select the desired tool, like `calculate_feature_score`.
3. Provide the necessary input, such as the feature ID or content to check.
4. The MCP runs the WCAG diagnostic and returns a detailed, actionable report.

## Frequently Asked Questions

**Does this MCP only check for WCAG AA compliance?**
The MCP analyzes compliance against WCAG standards. You can run checks to identify specific technical failures against the selected WCAG standard.

**What kind of output do I get from the feature score?**
The output is a high-level health assessment score, plus a list of specific compliance gaps that need attention.

**Can I check if my AI-generated alt text is good enough?**
Yes. The `analyze_content_description` tool evaluates the quality of text descriptions for visual assets, telling you if they are sufficient.

**What is a 'focus trap' in this context?**
A focus trap is a navigation issue where a user, using only a keyboard, gets stuck in a part of the interface and cannot tab out of it. The `check_navigation_flow` tool detects these.

**Is this tool for general accessibility or specific AI features?**
It is specialized for AI-driven software features. It provides diagnostic tools to evaluate the accessibility of AI interactions, not just static web pages.