| name | aso |
| description | Make a site agent-ready using Cloudflare AI Search and the isitagentready.com checklist. Audits the current site, then implements: Cloudflare AI Search (hybrid vector + keyword), markdown content negotiation (acceptmarkdown.com), robots.txt AI rules, sitemap, MCP Server Card, and Agent Skills discovery. Use when asked to add search, make a site agent-ready, or optimize for AI agents. |
ASO
You are making a site fully agent-ready using Cloudflare AI Search and the standards defined at isitagentready.com. Work through each phase in order.
Target: {{args}}
Phase 1: Detect Current State
Note: The shell snippets in this skill assume a POSIX shell. On Windows, run them via Git Bash or the Bash tool rather than PowerShell.
Run silently before asking anything. Check what's already in place:
ls package.json wrangler.jsonc wrangler.toml next.config.* astro.config.* 2>/dev/null
cat package.json 2>/dev/null | grep -E '"(next|astro|remix|nuxt|svelte|hono|workers-sdk|wrangler)"'
cat wrangler.jsonc 2>/dev/null || cat wrangler.toml 2>/dev/null
cat public/robots.txt 2>/dev/null || cat static/robots.txt 2>/dev/null
ls public/sitemap*.xml static/sitemap*.xml 2>/dev/null
ls public/.well-known/ 2>/dev/null
cat public/.well-known/mcp.json 2>/dev/null
cat public/.well-known/ai-plugin.json 2>/dev/null
grep -r "ai_search" wrangler.jsonc wrangler.toml 2>/dev/null
grep -r "text/markdown" src/ app/ pages/ 2>/dev/null | head -10
Note: Is this a Workers project? What framework? What's already done vs. missing?
Phase 2: Audit Against isitagentready.com
Check each category and mark as DONE / MISSING:
Discoverability
- robots.txt with AI bot rules and Sitemap reference
- XML sitemap
- Link response headers (
Link: <URL>; rel="alternate"; type="text/markdown")
Content Accessibility
- Markdown content negotiation (
Accept: text/markdown → respond with .md)
Bot Access Control
- AI bot rules in robots.txt (GPTBot, ClaudeBot, PerplexityBot, etc.)
- Web Bot Auth (if needed for authenticated content)
Protocol Discovery
- MCP Server Card at
/.well-known/mcp.json
- Agent Skills listing
- API Catalog (if the site has an API)
Search
- Cloudflare AI Search instance with hybrid indexing
Show the user a scored summary before proceeding:
Discoverability: [2/3 done]
Content: [0/1 done]
Bot access: [1/2 done]
Protocol: [0/2 done]
Search: [0/1 done]
Ask: "Which of these do you want me to implement? I can do all of them, or we can start with the highest-impact ones first." Default recommendation: all.
Phase 3: Cloudflare AI Search Setup
3a. Install Wrangler (if not already installed)
command -v bun >/dev/null 2>&1 && PM=bun || (command -v pnpm >/dev/null 2>&1 && PM=pnpm || PM=npm)
$PM add -D wrangler
3b. Create a search instance
npx wrangler ai-search create <project-name>-search
For binding config, Worker code, indexing, and hybrid search setup, see references/search-setup.md
Phase 4: Markdown Content Negotiation
AI agents save tokens when they receive Markdown instead of HTML. Implement Accept: text/markdown content negotiation per acceptmarkdown.com.
For framework-specific implementation, see references/markdown-negotiation.md
Phase 5: robots.txt AI Rules
Add AI bot directives alongside a sitemap reference, explicitly allowing or blocking AI crawlers as needed.
See references/robots-sitemap.md for the full robots.txt templates
Phase 6: XML Sitemap
Generate and serve a sitemap at /sitemap.xml; reference it in robots.txt.
See references/robots-sitemap.md for framework-specific sitemap generation
Phase 7: MCP Server Card
Publish at /.well-known/mcp.json so AI agents can discover your MCP capabilities.
See references/discovery-files.md for the mcp.json template
Phase 8: Agent Skills Listing
Publish at /.well-known/agent-skills.json to advertise what actions agents can take on your site.
See references/discovery-files.md for the agent-skills.json template
Phase 9: Verification
Run verification checks and confirm each passes — see references/verification.md
Completion Checklist
