squirrelscan
GitHub

Answer blocks: can an answer engine quote this page

Question headings with no short, self-contained answer, and data tables or charts with no source. squirrel flags both on content pages and never lowers the score.

What answer blocks are

An answer engine quotes a page by lifting a passage. The passages that lift cleanly are the ones that stand alone: a question as a heading, then an answer in the first sentence or two that makes sense without the text around it. A passage that opens with “It is a file that…” loses its subject the moment it is lifted. A table or chart with no source next to it is hard to attribute, so there is less reason to quote it.

This rule is advisory. It reports with the info severity, so it never changes the health score or the Agents score. Topic and authority matter more than format for being cited, so treat a finding as a suggestion to tighten the copy, not as a defect.

What squirrel checks

The rule runs on every crawled page and emits up to two checks. A page with no question headings, data tables or charts emits one skipped answer-blocks check with the message No question headings, data tables or charts on this page. Product, category, search, contact, profile, event, local and media pages are skipped with Not a content page, answer blocks not assessed, along with soft 404 pages and pages kept out of the index with noindex.

  • answer-blocks: looks at every h2, h3 and h4 whose text ends in a question mark and has at least three words. Headings inside nav, footer, aside or hidden elements are ignored. The answer is the first paragraph after the heading, before the next heading. A list, table, code block or quote straight after the heading counts as a direct answer. The check warns with Question headings without a short, self-contained answer in the first two sentences when a question heading has no text under it, when the first sentence opens with a pronoun (it, they, he, she, them, their, or this, that, these, those standing in for a noun, as in This is...), or when the first sentence is longer than 60 words. Each flagged heading is an item, labelled with the reason. Otherwise it passes with Every question heading is followed by a direct answer.
  • figure-sources: looks at data tables (three or more rows, not laid out with role="presentation") and figures that hold a chart, an svg, a canvas or a table. A source counts when the caption, figcaption or tfoot, or the element directly before or after (up to 400 characters), says source, credit, via, according to, courtesy or the like, carries a copyright mark, or links to an external page. The check warns with Data tables or charts with no source named next to them and lists each one, or passes with Every data table and chart names its source.

The heading and source tests are text patterns, so they are heuristics. A heading that ends in a question mark but is not a question, or an answer that starts with a pronoun that does have a clear subject, can be flagged.

How to fix it

Put the answer under the question, and make it stand alone:

<h2>What is a sitemap?</h2>
<p>A sitemap is an XML file that lists the URLs of a site. Crawlers read it to find pages they would not reach by links alone.</p>

Name the subject instead of opening with “It” or “This”, and keep the first sentence short. Put the detail after it. For a table or chart, add a caption that names the source:

<table>
  <caption>Monthly users. Source: <a href="https://example.org/report">Example Research, 2026</a></caption>
  ...
</table>

Citability in the Agents section

The signals an answer engine weighs for trust and recency already have their own rules: eeat/content-dates, content/freshness, content/date-agreement, eeat/author-byline, eeat/author-expertise and eeat/citations. The Agents section of the console, text, markdown and HTML reports prints one line that links to each of them with its status (ok, issues with the rule id, or not checked). The findings are not repeated there, and the line never touches the score.

Rule IDax/answer-blocks
CategoryAgent Experience
ScopePer-page
Severityinfo
Weight1/10

Enable / disable

Disable this rule

[rules]
disable = ["ax/answer-blocks"]

Disable all Agent Experience rules

[rules]
disable = ["ax/*"]

Enable only this rule

[rules]
enable = ["ax/answer-blocks"]

enable replaces the default ["*"], so this is the only rule selected. Leave disable out: a rule that matches any disable pattern is skipped even when enable names it. Only an enabled setting under a rule’s own [rule_options] table overrides both lists.

Check names by outcome

Each check name this rule can emit, by what it means to a tool comparing two audits. A problem check is a finding. A clean check says the rule ran over what it covers and found nothing; it clears the problem checks listed beside it, so a finding is fixed when the rule ran over the same scope and a clean check took its place. A not checked check says the rule did not evaluate, and never clears a finding.

Check nameOutcomeClears
answer-blocksProblem
answer-blocksCleananswer-blocks
figure-sourcesProblem
figure-sourcesCleanfigure-sources
answer-blocksNot checked

Agent experience findings ship in every audit next to the SEO, performance and security rules. See Generative engine optimization for how an agent works through a report.

Check your site

Run squirrel audit https://example.com and open the Agent Experience section of the report. Signed-out local audits need no account and cost nothing. Signed-in audits spend credits.

Type to search…

↑↓ navigate↵ openesc close