🤖 SEO Testing with GitHub Copilot

Integrate SEODiff evaluation into GitHub Copilot agent workflows.

GitHub Copilot’s agent mode can call external APIs via MCP servers or custom instructions. Configure it to validate SEO quality on pages you’re editing — using SEODiff’s agentic evaluation endpoint.

System Prompt / Configuration

Copy this prompt and paste it into your IDE’s AI configuration:

# SEO Quality Instructions for GitHub Copilot
# Add to .github/copilot-instructions.md in your repository

## SEO Evaluation Workflow

When editing HTML pages, React components, or any template that generates web pages:

1. After making changes, evaluate SEO quality using the SEODiff API:

   curl -X POST "https://seodiff.io/api/v1/agent/evaluate" \
     -H "Authorization: Bearer $SEODIFF_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
       "urls": ["<page_url>"],
       "assertions": [
         {"rule": "has_h1"},
         {"rule": "has_schema"},
         {"rule": "no_placeholders"},
         {"rule": "max_js_ghost_ratio", "value": 0.1}
       ],
       "wait": true
     }'

2. Fix any failing assertions before marking the task complete.
3. Target an ACRI score of 70+ for all pages.

Setup

# Add to your repository's .github/copilot-instructions.md
# Copilot will follow these instructions in agent mode

Example Interaction

User: "Fix the blog post template — it's not showing rich results."

Copilot:
1. Evaluates current template: has_schema FAILED (no JSON-LD found)
2. Adds Article schema with headline, datePublished, author
3. Re-evaluates: All assertions pass, schema_coverage: 100%
4. "Added Article JSON-LD schema. ACRI improved from 52 to 71."