Agent readiness tools

Markdown for agents checker

Q: How do these checks count in the score?

markdown-negotiation is in the agent-conventions part: 30% of the score, and none of its checks holds a level. The methodology says what each state earns.

Enter a domain to request its homepage as markdown. The report opens on the type and the Vary header of the answer, with any change to make.

The Markdown for agents checker runs the full scan of 60 checks, not one request. The five exemplar scans of 20 October 2026 sent 29 to 89 requests each and took 3 to 8 seconds. A scan that reaches its 60 s deadline is reported as not measured.

A stored result: cloudflare.com, measured on 20 October 2026. Its report opens on the same panel yours will.

What is checked

The report opens on 1 check of rubric v0.9. Each is described here in the rubric's own words.

Markdown content negotiation

Some coding agents ask for text/markdown. Google Search does not use markdown.

markdown-negotiation is a scored check. Agent conventions: 30% of the score. A convention never holds a level.

The scanner requests the homepage again with Accept: text/markdown, text/html;q=0.5 and reads the status, Content-Type, and Vary of the answer. Present when the answer is text/markdown or text/x-markdown with Vary: Accept or Vary: *. Served with an error when markdown comes without Vary: Accept, when a markdown body comes as text/plain, or when the markdown request answers HTTP 400 or above, other than 429, while the homepage answers 2xx. Not used when the answer is HTML: the request accepts HTML, so that is a valid response. Also not used when the answer is another type that is not markdown, or has no Content-Type. Not measured when the request gets no HTTP response, when it answers HTTP 429, a rate limit on the scanner, or when it answers an error and the homepage does too. Google Search does not use markdown. Present earns the check's points in the agent-conventions part; Served with an error and Not used earn none.

Run the same measurement against your own site. Set the origin on the first line:

HOMEPAGE=https://example.com/
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/markdown, text/html;q=0.5' -D - -o /tmp/aws-md.body "$HOMEPAGE" | grep -iE '^(HTTP/|content-type|vary)'; head -c 300 /tmp/aws-md.body; echo

Read the markdown-negotiation check page for its sources and the change to make.

Questions about the Markdown for agents checker

Does Cloudflare's Markdown for Agents pass this check?

The check reads the answer, not what made it. The scanner requests the homepage again with Accept: text/markdown, text/html;q=0.5 and reads the status, Content-Type, and Vary of the answer. Present when the answer is text/markdown or text/x-markdown with Vary: Accept or Vary: *. Markdown for agents reads Cloudflare's documentation: the converted answer is text/markdown, and its Vary includes Accept. By the rule above, that answer is present. Next.js, WordPress, or an edge rule counts the same way.

Does Google Search use markdown?

Some coding agents ask for text/markdown. Google Search does not use markdown. Claude Code, OpenCode and Cursor send Accept: text/markdown by default (Cloudflare, 2026-04).

How do these checks count in the score?

markdown-negotiation is in the agent-conventions part: 30% of the score, and none of its checks holds a level. The methodology says what each state earns.

Why does it run the full scan?

The form starts the same scan as /scan, so the result is a scan record with a fingerprint you can verify. The report opens on markdown-negotiation, and the full report follows. A stored record younger than 6 hours opens instead of a new scan.