agent-ready.dev stops crediting catch-all pages as llms.txt
In short
agent-ready.dev changed S1, S10, S12, L1 and L10 so an HTML page returned with 200 at /llms.txt, /sitemap.md or /AGENTS.md no longer counts as the file. Sites with single-page-app catch-alls lose those checks; older scans were not re-scored. agent-ready.dev's methodology excludes checks marked unreliable by its JavaScript-rendering check, so one architectural choice does not penalise four dependent checks.
What happened
agent-ready.dev changed S1, S10, S12, L1 and L10 so an HTML page returned with 200 at /llms.txt, /sitemap.md or /AGENTS.md no longer counts as the file. Sites with single-page-app catch-alls lose those checks; older scans were not re-scored. 1
Date: 20 October 2026. Source type: a primary source (the company or standard body itself).
Who it involves
- agent-ready.devScanner that grades a site on 72 checks, 69 mapped to published specs (Vercel's agent readability guide, llmstxt.org and agent protocol specs), with an llms.txt sub-score and a separate MCP server grader.
Nearby items
More from the same source
- agent-ready.dev's methodology excludes checks marked unreliable by its JavaScript-rendering check, so one architectural choice does not penalise four dependent checks.1
Sources
- agent-ready.dev methodology. https://agent-ready.dev/methodology, read 2026-11-05.
Last reviewed 2026-11-05