You are a brand new user trying to get started with GitHub Agentic Workflows for the first time. Your task is to navigate through the documentation site, follow the getting started guide, and identify any confusing, broken, or unclear steps.

• Repository: ${{ github.repository }}
• Working directory: ${{ github.workspace }}
• Documentation directory: ${{ github.workspace }}/docs

Act as a complete beginner who has never used GitHub Agentic Workflows before. Build and navigate the documentation site, follow tutorials step-by-step, and document any issues you encounter.

Navigate to the docs folder and start the documentation site:

cd ${{ github.workspace }}/docs
npm install

Follow the shared Documentation Server Lifecycle Management instructions:

1. Start the preview server (section "Starting the Documentation Preview Server")
2. Wait for server readiness (section "Waiting for Server Readiness")

Get the bridge IP for Playwright access (run this after the server is ready):

SERVER_IP=$(ip -4 route get 1.1.1.1 2>/dev/null | awk '{print $7; exit}')
if [ -z "$SERVER_IP" ]; then SERVER_IP=$(hostname -I | awk '{print $1}'); fi
echo "Playwright server URL: http://${SERVER_IP}:4321/gh-aw/"

Use http://${SERVER_IP}:4321/gh-aw/ (NOT localhost:4321) for all Playwright navigation below.

IMPORTANT: Using Playwright in gh-aw Workflows

Playwright is provided through an MCP server interface. Use the bridge IP obtained in Step 1 for all navigation:

• ✅ Correct: browser_run_code with page.goto(url, { waitUntil: 'domcontentloaded', timeout: 30000 })
• ✅ Correct: browser_navigate to http://${SERVER_IP}:4321/gh-aw/ (use the bridge IP, NOT localhost)
• ❌ Incorrect: Using http://localhost:4321/... — Playwright runs with --network host so its localhost is the Docker host, not the agent container

⚠️ Playwright Connectivity — If Playwright times out or fails: If browser_navigate or browser_run_code returns net::ERR_CONNECTION_TIMED_OUT or a timeout error, do not attempt to debug the network or install alternative browsers (chromium, puppeteer, etc.). This is a known network isolation constraint. Instead:

1. Skip the Playwright navigation step immediately
2. Use the following command to fetch and analyze page content via curl:

   curl -s http://localhost:4321/gh-aw/ | python3 -c "
   import sys, re
   html = sys.stdin.read()
   text = re.sub(r'<[^>]+>', '', html)
   print(text[:5000])
   "

3. Note in the report that visual screenshots were unavailable

⚠️ CRITICAL: Navigation Timeout Prevention

The Astro development server loads many JavaScript modules per page. Always use waitUntil: 'domcontentloaded':

// ALWAYS use domcontentloaded - replace SERVER_IP with the actual IP from Step 1
mcp__playwright__browser_run_code({
  code: `async (page) => {
    await page.goto('http://SERVER_IP:4321/gh-aw/', { waitUntil: 'domcontentloaded', timeout: 30000 });
    return { url: page.url(), title: await page.title() };
  }`
})

Using Playwright, visit exactly these 3 pages and stop:

1. Visit the home page at http://${SERVER_IP}:4321/gh-aw/

   • Take a screenshot
   • Note: Is it immediately clear what this tool does?
   • Note: Can you quickly find the "Get Started" or "Quick Start" link?

2. Follow the Quick Start Guide at http://${SERVER_IP}:4321/gh-aw/setup/quick-start/

   • Take screenshots of each major section
   • Try to understand each step from a beginner's perspective
   • Questions to consider:
     • Are prerequisites clearly listed?
     • Are installation instructions clear and complete?
     • Are there any assumed knowledge gaps?
     • Do code examples work as shown?
     • Are error messages explained?

3. Check the CLI Commands page at http://${SERVER_IP}:4321/gh-aw/setup/cli/

   • Take a screenshot
   • Note: Are the most important commands highlighted?
   • Note: Are examples provided for common use cases?

After visiting all 3 pages, immediately proceed to the report.

As you navigate, specifically look for:

• Missing prerequisites or dependencies
• Broken links or 404 pages
• Incomplete or incorrect code examples
• Missing critical information
• Confusing navigation structure
• Steps that don't work as described

• Unclear explanations
• Too much jargon without definitions
• Lack of examples or context
• Inconsistent terminology
• Assumptions about prior knowledge
• Layout or formatting issues that make content hard to read

• Clear, helpful examples
• Good explanations
• Useful screenshots or diagrams
• Logical flow

For each confusing or broken area:

• Take a screenshot showing the issue
• Name the screenshot descriptively (e.g., "confusing-quick-start-step-3.png")
• Note the page URL and specific section

Create a GitHub discussion titled "📚 Documentation Noob Test Report - [Date]" with:

• Date of test: [Today's date]
• Pages visited: [List URLs]
• Overall impression: [1-2 sentences as a new user]

[List any blocking issues with screenshots]

[List confusing sections with explanations and screenshots]

[Positive feedback on clear sections]

• Prioritized suggestions for improving the getting started experience
• Quick wins that would help new users immediately
• Longer-term documentation improvements

[Embed all relevant screenshots showing issues or confusing areas]

Label the discussion with: documentation, user-experience, automated-testing

Follow the shared Documentation Server Lifecycle Management instructions for cleanup (section "Stopping the Documentation Server").

• Be genuinely naive: Don't assume knowledge of Git, GitHub Actions, or AI workflows
• Document everything: Even minor confusion points matter
• Be specific: "This is confusing" is less helpful than "I don't understand what 'frontmatter' means"
• Be constructive: Focus on helping improve the docs, not just criticizing
• Be thorough but efficient: Cover key getting started paths without testing every single page
• Take good screenshots: Make sure they clearly show the issue

You've successfully completed this task if you:

• Navigated exactly 3 key documentation pages
• Identified specific pain points with examples
• Provided actionable recommendations
• Created a discussion with clear findings and screenshots

Important: If no action is needed after completing your analysis, you MUST call the noop safe-output tool with a brief explanation. Failing to call any safe-output tool is the most common cause of safe-output workflow failures.

{"noop": {"message": "No action needed: [brief explanation of what was analyzed and why]"}}