Scoring Rubric — Five Pillars, 100 Points

The rubric scores every property in five pillars summing to 100 points.

Pillar	Weight	Measures	Owner
1. Crawl & Discovery	20 pts	robots.txt with AI-bot rules, sitemap validity, canonicals, Link headers, no-JS discoverability	Engineering
2. Content Legibility	25 pts	Markdown content negotiation and URL-suffix fallback, SSR, token-efficient output, llms.txt, semantic HTML	Engineering + Content
3. Answer-Engine Quality	30 pts	Entity clarity, E-E-A-T, answer-first structure, FAQ / definition / comparison patterns, freshness governance, primary-source citations (Framix-only — not in Cloudflare’s rubric)	Content + SEO
4. Bot Access & Trust	10 pts	Content Signals in robots.txt, AI-bot governance, Web Bot Auth verification (when site originates bot traffic), AI-crawler analytics	Engineering
5. Capabilities & Commerce	15 pts	API Catalog (RFC 9727), OAuth metadata (RFC 8414/9728), MCP Server Card (SEP-2127), Agent Skills index, A2A Agent Card, x402 payment flows	Engineering (conditional)

Re-normalisation by property type

Pillar 5 drops to N/A on pure-content sites with no APIs or transactional surface; weights re-normalise across pillars 1–4. On agent-targeted SaaS or commerce properties, Pillar 5 weight increases to 25 pts, drawn from Pillar 3.

How this maps to Cloudflare’s IsItAgentReady

Cloudflare category	Framix pillar	Notes
Discoverability	1	Same scope
Content Accessibility	2	Same scope
Bot Access Control	4	Cloudflare adds Web Bot Auth as a check
Protocol Discovery	5	MCP Server Card, Agent Skills, WebMCP, API Catalog, OAuth
Commerce	5	x402, MPP, UCP, ACP — tracked in Cloudflare, not yet scored
(none)	3	Framix-only. Authority, citations, and answer structure.

The five pillars map onto the two-layer model — Pillars 1, 2, 4, 5 measure Layer A (Agent Readiness); Pillar 3 alone measures Layer B (GEO Visibility).

Item-level checklist in Universal Requirements.