Agent-ready website checklist
Use this implementation checklist to improve the public signals Agent Web Check can verify today: discovery, content readability, access policy, protocol readiness, schema, and shallow task completion.
Discovery pass
Confirm HTTPS, robots.txt, sitemap.xml, canonical tags, crawlable homepage text, readable titles and descriptions, and useful internal links to product, docs, pricing, contact, and policy pages.
Content pass
Publish /llms.txt, keep important copy out of JS-only states, use readable headings, and expose markdown-friendly alternatives for docs or guides when they materially improve parsing.
Policy pass
Review robots.txt for broad crawler blocks, explicit AI crawler rules, conflicting directives, sitemap references, WAF behavior, and public fetch responses that accidentally look like access denial.
Protocol pass
Add relevant well-known metadata, API catalogs, OpenAPI for API products, OAuth metadata for authenticated APIs, and WebMCP or MCP discovery where repeatable workflows actually exist.
Task pass
Make CTAs semantic, label forms, expose pricing and policy pages, provide clear confirmation or fallback contact paths, and make checkout or booking paths obvious when those workflows are public.
Do not try to fix every signal at once. Start with public discovery and content readability, then add protocol files only where they describe real workflows.
- Where should I start if the score is low?
- Start with public discovery and readable content: HTTPS, robots.txt, sitemap.xml, crawlable homepage copy, headings, metadata, and /llms.txt for important public pages.
- How often should I recheck a site?
- Recheck after launches, docs migrations, pricing changes, policy updates, new API releases, or changes to crawler access controls because those updates can alter public agent evidence.
- Should I fix every protocol signal first?
- No. Fix the signals that match real workflows. Content, policy, schema, and task clarity often matter before adding advanced protocol discovery.