Methodology & Observed Failure Modes
Six weighted technical checks grounded in live telemetry
Every score comes from verifiable HTTP traces, Chromium DOM diffs, open registry APIs, and web-grounded AI queries. Nothing is simulated.
01 · Crawler Access
7/11 Blocked
RFC 9309 + Targeted vs. Site-Wide WAF
Separates intentional robots.txt opt-outs from silent firewall blocks. On infinifi.xyz and loopscale.com, Cloudflare drops all 14 bots; on axis.to, robots.txt line 143 allows PerplexityBot while Cloudflare drops it with a 25-byte HTTP 403.
02 · JS-Off Rendering
4/11 CSR Blackouts
Headless Chromium JS-On vs. JS-Off Diff
Captures what non-JS crawlers extract vs. human browsers. current.finance drops from 494 rendered words to 0 raw HTML words, prjx.com from 964 to 7, and mezo.org sub-routes /docs and /earn/lock drop to 0.
03 · Schema & Graph
Canonical Leaks
JSON-LD Entities & Subdomain Consistency
Validates Organization, WebSite, and FAQPage in raw HTML. Catches missing root schemas on nado.xyz, mezo.org, and loopscale.com, plus multi-host splits across entropy.io, pro.entropy.io, and entropy.trade.
04 · Machine Corpus
7/11 Missing llms.txt
Sitemap Hygiene & /llms.txt Endpoint DNS
Verifies markdown validity and tests referenced URLs and RPC hosts live. Catches nado.xyz leaking /emailtest in sitemap.xml, prjx.com serving HTML on /llms.txt, and monad.xyz/llms.txt pointing to unresolved rpc.testnet.monad.xyz.
05 · Entity Footprint
Off-Site Trust
DefiLlama, CoinGecko, GitHub & Wikidata
Audits third-party truth sources that LLMs cross-reference for safety and disambiguation. Prevents unverified aggregators like tradingstrategy.ai or iq.wiki from defining your protocol risk and token identity.
06 · Live Citations
0/44 Category Hits
Multi-Engine Web-Grounded Query Benchmarking
Measures brand, comparison, and category share across Perplexity Search and Google Gemini grounding. On Monad (86/100 A), Gemini still hedged between 400ms and 300ms block times because official AI docs drifted (v0.16.1 vs. v0.16.5).