Solo founder here. The "blue link" layer of search is being eaten by the answer layer — Google AI Overviews, ChatGPT browse, Claude artifacts, Perplexity citations. They rely on a set of signals most sites haven't shipped: llms.txt, schema.org JSON-LD, agent-card.json (A2A), MCP server-card, OpenAPI, bot-friendly robots.txt.
I scanned ~200 sites this month: 96% fail at least 6 of the 12 most basic signals. Most fail all 12.
Built and open-sourced a CLI scanner:
npx agentfix-mini-scanner yoursite.com
12 of 33 signals. MIT. Zero deps. Node 18+. ~5 sec. Grade A–F + concrete fixes per check.
There's also a free interactive iceberg viz: workspace.agentfix.pro — what Google sees ↑ what AI agents reach for ↓ what nobody ships yet 🌊
Full product at agentfix.pro = 33 signals + ZIP fix-pack generated from your real content. One-time $1/$29/$99 (no subscription).
Repo: https://github.com/Virt92/agentfix-oss-scanner
Scanner: npx agentfix-mini-scanner yoursite.com
Workspace viz: https://agentfix.pro
Top comments (0)