AI readiness report

www.shidianguwen.com

7/15

Invisible

7 of 15 checks passed · Standard v1.2 · Engine 0.1.0-m0

Hard gate failed. At least one group A check failed, so the verdict is capped at “Invisible” however well the rest of the page scores. Group A asks whether an AI engine can reach the page at all.

Positioning: Chinese-language domestic site (domestic). The hard gate depends on this: an English-language site is measured on the international AI crawlers, a Chinese-language site on the China ecosystem.

Fix in this order

  1. A4 No dead-link pollution A failed group A check caps your grade at Invisible however well everything else scores, so this one unlocks the grade.
  2. C3 There is original data or first-hand information 93% of the sites we have measured fail this check — it is the common case, not an edge case.
  3. B2 Structured data (JSON-LD) is present 86% of the sites we have measured fail this check — it is the common case, not an edge case.
  4. C4 There is a visible last-updated date 86% of the sites we have measured fail this check — it is the common case, not an edge case.
  5. B4 An llms.txt is present 79% of the sites we have measured fail this check — it is the common case, not an edge case.
  6. B3 Heading hierarchy is semantic 50% of the sites we have measured fail this check — it is the common case, not an edge case.

A. Can AI reach you?

Can an AI engine reach the page at all?

  1. A4 No dead-link pollution Fail

    1 of 20 checked internal link(s) failed (5.0%), above the 1% threshold.

    How to fix it. One of the 20 checked internal links fails; find that link and either fix its target URL or remove it. Then re-scan so the failure rate drops below the 1% threshold.tailored to this site

  2. A1 robots.txt allows the AI crawlers Pass

    All 5 China-ecosystem crawlers may fetch the site (0 declared explicitly). Positioning: domestic.

  3. A2 Content is in the HTML, not only after JavaScript Pass

    The raw HTML already contains 8185 characters of readable text, so a crawler that does not run JavaScript can read the page.

  4. A3 A sitemap exists Pass

    The sitemap is valid XML and lists 169 URL(s); first entry: https://www.shidianguwen.com/sitemap-category.xml.

  5. A5 Reachable and HTTPS is valid Pass

    The homepage answers HTTP 200 over HTTPS and certificate valid for 63 more day(s); no mixed content found.

B. Can AI understand who you are?

Can it tell who you are and what the page is about?

  1. B2 Structured data (JSON-LD) is present Fail

    No application/ld+json block was found.

    How to fix it. No application/ld+json block exists on the page. Add a JSON-LD block describing the site and its main content type so the page carries machine-readable structured data.tailored to this site

  2. B3 Heading hierarchy is semantic Fail

    Problems: no <h1>. Heading sequence: [3, 4, 4, 4, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2].

    How to fix it. The page has no <h1> and its headings run h3/h4 down into a long run of h2s. Add one <h1> for the page's main topic and rebuild the sequence so levels descend in order without skipping.tailored to this site

  3. B4 An llms.txt is present Fail

    No llms.txt was found at the site root.

    How to fix it. No llms.txt is served at the site root. Add an llms.txt file at the domain root that points AI crawlers to the key pages and describes what the site covers.tailored to this site

  4. B5 The operator behind the site is verifiable Fail

    Only 1 trust signal(s) found: a copyright line. Anonymous sites are rarely treated as citable sources.

    How to fix it. Only one trust signal, a copyright line, was found, leaving the operator unverifiable. Add named operator information, such as an about page and contact or registration details, so the site can be attributed to a real entity.tailored to this site

    This rule cannot decide reliably on its own (confidence 80%), so treat it as a suggestion rather than a verdict.

  5. B1 Title and description are complete Pass

    Title (15 chars): "诗典古文分享 - 诗典古文分享"; description (109 chars) present.

C. Will AI choose to cite you?

Will it choose to quote you rather than someone else?

  1. C1 The conclusion comes first Fail

    The first paragraph is only 4 characters: "诗典古文".

    How to fix it. The opening paragraph is only 4 characters ("诗典古文"), so no answer is stated up front. Replace it with a complete sentence that states the page's main conclusion before any supporting content.tailored to this site

    This rule cannot decide reliably on its own (confidence 70%), so treat it as a suggestion rather than a verdict.

  2. C3 There is original data or first-hand information Fail

    Only 0 percentage figure(s) and 0 data point(s) were found — mostly generic prose.

    How to fix it. The page contains no percentage figures and no data points, only generic prose. Add first-hand data or original figures that only this site could publish, such as counts from the content it hosts.tailored to this site

    This rule cannot decide reliably on its own (confidence 60%), so treat it as a suggestion rather than a verdict.

  3. C4 There is a visible last-updated date Fail

    No updated date was found in the page text or metadata.

    How to fix it. No updated date appears in the page text or metadata. Add a visible last-updated date on the page and mirror it in the page metadata.tailored to this site

  4. C2 There are extractable structure blocks Pass

    Found 9 list(s) with 3+ items — these can be lifted out of the page as a self-contained answer.

  5. C5 There is third-party reference evidence Pass

    31 distinct external domain(s) referenced.

Analysis · written from the measurements above

Crawling and reading are not the problem — the content layer is: the page has no <h1>, no JSON-LD, no llms.txt, no visible last-updated date, and only one trust signal (a copyright line), and its opening paragraph is just four characters long. Start by adding a single <h1> and a JSON-LD block, then fix the one internal link that failed (1 of 20 checked), since those are concrete and quick. After that, add an operator identity and a visible last-updated date so the site can be treated as a citable source.

Score history

One measurement so far. Apply the fixes above, then recheck to prove the score moved.

1 measurement, 15 checks each.
MeasuredScoreChangeFixedNewly failing
23 Sep, 08:40 First measurement 7 / 15 — — —

Next step

Apply the failing checks above in group order — group A first, because a site that cannot be reached or read cannot be cited. Then recheck to see the score move.

Download the repair pack

The pack contains robots.txt, llms.txt and a JSON-LD block, plus a deploy script that is a dry run by default and backs up anything it replaces, and a verify script that confirms the deploy landed.

Report generated 2026-09-23T08:40:47 (UTC).

Check a site