Get quoted by answer engines and clicked in search.
Audits a page or site on both tracks and returns effort-vs-impact ranked fixes so an AI answer engine will quote you.
Inside: Entity clarity, liftable answers, schema, internal links, and a prioritized fix list.
Install it in one line, or paste it in.
~/.claude/skills/ and runs automatically when it is relevant.Connect your context. Set it to your motion.
it sees the queries you already rank near, so the fixes target winnable ground, not fantasy keywords.
it sees which pages earn attention, so you fix the pages that can move the number.
it can check schema and internal links across templates, not one page at a time.
This was built for a content site that wants to be quoted by answer engines. Set these to your stack:
Audit against the one question the page should win. A page that answers ten questions ranks for none. Run it again per question.
| Set this | What it is | Example |
|---|---|---|
| TARGET_QUESTION | the query the page should own | "how do I export to CSV" |
| ENTITY | the thing the page is about | your producta concepta place |
| SCHEMA types | structured data to check | FAQPageHowToArticleProduct |
| SEARCH_CONSOLE | your query-data connector | a search console |
| ANALYTICS | your web-analytics connector | a web-analytics tool |
| CMS | where the page is edited | a CMS |
| FRESHNESS_DAYS | age at which content is stale | 180 (re-tune to your topic) |
Everything the skill does, in full.
Reads a page (or a set of pages) and scores it two ways at once: can a classic search engine rank it, and can an answer engine lift a clean quote from it. Then it returns a ranked fix list, each fix with the reason it matters and the effort it takes, so you work the cheap high-impact items first. It is written for the world where half your traffic never clicks, because the answer got read aloud.
- 1Entity clarity
An answer engine has to know what the page is about in one pass. The skill checks that the ENTITY is named plainly, defined early, and consistent. Vague pages do not get quoted.
- 2Structured answers
Answer engines lift the cleanest self-contained passage. The skill checks that the TARGET_QUESTION is answered directly, high on the page, in a liftable block: a short definition, a list, or a labeled step sequence. If the answer is buried under a story, it flags it.
- 3Schema
Checks whether the right SCHEMA type is present and valid for the page's job (an FAQ block, a how-to sequence, an article byline and date). Missing or malformed schema is a named fix, not a vague "add structured data."
- 4Internal links
Checks that the page is linked to from related pages with descriptive text, and that it links out to its own supporting pages. An orphan page is invisible to both kinds of engine.
- 5Freshness
Flags content older than FRESHNESS_DAYS with no update, and checks that a real published-or-updated date is visible. Answer engines discount stale pages on time-sensitive topics.
- 6Prioritized fix list
Every finding is scored effort against impact and sorted. Cheap high-impact fixes go to the top. You get a worklist, not a wall of red.
- No claimed ranking or position without a connected query source. Unknown stays unknown.
- Every fix names the specific element to change, never "improve your content."
- Illustrative traffic or position numbers are marked as examples, never presented as your real data.
SEO + AEO AUDIT · page: /guide/csv-export · question: "how do I export to CSV"
Classic search: 66 / 100
Answer-engine: 41 / 100 (the gap is why you are not quoted)
Top fixes (effort -> impact):
1. LOW -> HIGH Add a 2-sentence direct answer above the fold.
Right now the answer starts in paragraph 5.
2. LOW -> HIGH Add HowTo schema to the numbered steps.
3. MED -> HIGH Link this page from the 3 related feature pages.
It is currently an orphan.
4. LOW -> MED Add a visible "updated" date. Last touch: unknown.
5. MED -> MED Break the wall of text into labeled steps.
Next: connect a search console to target queries you already rank near.
Where an operator takes this next.
The one-page audit proves the gap is real. Here is the version that watches the whole site.
The fix list gets sharper every run instead of starting from scratch each time.
Run this as a scheduled Claude task across your CMS templates so every page gets re-scored monthly instead of one at a time by request.
Connect Google Search Console so the fix list points at questions you're already close to ranking for, not guesses.
Log each run's score to a sheet or Amplitude so you can see whether the gap between classic and answer-engine visibility is closing.
One skill is the on-ramp.
A single skill does one job. Chained into a playbook, or run as a full build, it becomes a system. Here is where this one plugs in.