A11y Demo Site Verified accessibility pipeline.

Evidence Generation

Scans HTML for WCAG violations and outputs structured findings with JSON Pointers to the offending elements.

Cryptographic Provenance

Every finding is signed with a SHA-256 digest. Verification proves evidence has not been tampered with since scan time.

Fix-Oriented Advisories

a11y-assist converts raw findings into actionable tasks with code examples showing exactly how to fix each issue.

CI/CD Integration

GitHub Actions workflows demonstrate blocking builds on regressions and uploading results as downloadable artifacts.

Intentional Bugs

The demo HTML contains real WCAG violations (missing lang, alt, labels) so you can see the full pipeline in action.

Artifact Inspection

Download CI artifacts to inspect findings, provenance records, digest chains, and advisory outputs locally.

Prerequisites

# Node.js 20+ and Python 3.10+
npm install -g a11y-evidence-engine
pip install a11y-assist

Run locally

# Clone and run the pipeline
git clone https://github.com/mcp-tool-shop-org/a11y-demo-site
cd a11y-demo-site
./scripts/a11y.sh

What happens

# 1. Scans html/ directory
# 2. Generates findings in results/
# 3. Ingests + verifies provenance
# 4. Outputs summary to console

Inspect results

# Start with the summary
cat results/a11y-assist/ingest-summary.json

# Then the advisories
cat results/a11y-assist/advisories.json

1. Scan

a11y-evidence-engine scans HTML files and emits structured findings with element selectors and rule references.

2. Sign

Each finding gets a provenance envelope: a canonical JSON representation hashed with SHA-256 and stored alongside the evidence.

3. Verify

a11y-assist recomputes the digest from the evidence and compares it to the stored hash. Match = VERIFIED.

4. Advise

Verified findings are converted into fix-oriented advisories with code snippets, grouped by rule and severity.