agentleFS
peakanswerpeakanswer/peak-answer-skills/skills/answer-gap-content/SKILL.md

answer-gap-content

Skill0 starsChanged 49 days agoMITView on GitHub

Decide what to write when an AI engine recommends a competitor instead of you, and brief the piece. Use when someone has a list of questions they lose, asks what content to write for AI search, or wants a content plan aimed at getting cited.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write for one question per page.

    line 30
  2. Distinctive:

    The answer, in one sentence: If you cannot write it, the page is not ready.

    line 37
  3. Distinctive:

    If the page is not picked up at all, the likeliest causes in order are: it is not indexed, the site blocks a crawler, the question was not …

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Answer gap content

You know a question where an engine names a competitor and not you. This is how to decide whether content fixes it, and what to write if so.

First, check content is the answer

Three causes, one fix each. Writing is the answer to one of them.

What you findThe fix
The engine cannot fetch or parse the siteCrawler access, server rendering. geo-audit skill
The site has no page that answers the questionWrite it. This skill
The pages the engine cites do not mention youGet into those pages. Below

Check the citations before writing anything. If the engine answered that question out of three listicles and a forum thread, a new page on your own domain is unlikely to displace them in the short term, and the work is outreach to those four pages instead. Writing is the default recommendation in this field and it is frequently the wrong one.

Choosing what to write

Rank candidate pieces by three things, in this order:

  1. Does the question have commercial intent? Someone choosing a tool beats someone learning a definition, every time.
  2. Is the answer contested? A question where four different brands win across runs is winnable. One where the same incumbent wins every run, on their own comparison page, is not, and the effort belongs elsewhere.
  3. Can you say something first-hand? A page that restates the same five bullets as the pages already cited gives a model no reason to switch. Your own numbers, your own test, your own product's behaviour, is the reason.

Write for one question per page. A page covering six questions is beaten on all six by six pages.

The brief

For each piece, fix these before writing:

  • The exact question, as a buyer types it. Not a keyword.
  • The answer, in one sentence. If you cannot write it, the page is not ready. This sentence opens the page.
  • The pages currently cited for that question, and what each is missing. That gap is the page's reason to exist.
  • The proof: the number, test, dataset or product behaviour only you can state.
  • The liftable block: the comparison table, the FAQ, the procedure. See answer-readiness.
  • Who is named. Including competitors. A comparison page that only names you is not a comparison, and models notice.

Writing it so it gets quoted

The shape rules live in answer-readiness; the short version is answer first, headings that match questions, something liftable in the middle, and an attributable claim. Beyond shape:

  • Name competitors honestly, including where they are the better fit. Pages that engines quote in comparison answers are the ones that read as evaluations rather than as pitches. A page that says a rival wins on one axis is more quotable, not less.
  • Be specific enough to be wrong. "Most teams see faster onboarding" cannot be cited. "In our 69-question test across 14 niches, 42% of answers named no vendor at all" can.
  • Date it and keep it dated. Engines discount undated pages in fast-moving categories, and this is one.
  • One page per question, updated, not a new page each quarter. Three near-duplicate posts split whatever authority the topic earns.

After it ships

Give it two to six weeks before judging. Then re-ask the same question, in the same engines, the same number of times, and check three things: is the brand named now, is the new page cited, and did the previously cited pages change. A page that is cited but does not move presence is usually a positioning problem, not a content one.

If the page is not picked up at all, the likeliest causes in order are: it is not indexed, the site blocks a crawler, the question was not contested in the first place, or the page says nothing the cited pages do not.

With the Peak Answer MCP connected

my_questions with only_losing gives the gap list with the winner named, and my_visibility gives the measured before-and-after rather than an impression of one. generate with content_brief produces a brief in the shape above for a specific question, and with faq_schema produces valid FAQPage JSON-LD from a page's real questions.

Free, no signup: https://peakanswer.com/tools/prompt-gap-checker

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.