- Home
- Skills
- Marketing & SEO
- GEO Citation Playbook
Works with the AI Agents you already use
GEO Citation Playbook
This skill finds out *why* for your page and tells your agent exactly what to change.
$3.50
GEO Citation Playbook
Example session with this skill installed
Audit https://example.com/best-coffee-makers for visibility in ChatGPT and Google AI Overviews. I can edit the site in WordPress. Focus on why we aren't cited in 'best for home office' queries.
- Read your context and instructions
- Compiled the geo citation playbook
I have completed the GEO audit for your coffee maker guide. The primary blockers are a restrictive robots.txt directive and a lack of structured data for product specifications.
GEO Audit & Fix Plan: Best Coffee Makers
| Issue | Impact | Evidence Tier | Recommended Action |
|---|---|---|---|
| Crawl Block | Critical | A | Update robots.txt to allow GPTBot and Google-Extended. Currently blocked. |
| Data Extraction | High | B | Add Product Schema with offers and aggregateRating. |
| Claim Specificity | Medium | B | Replace vague adjectives with verified technical specs (e.g., "fast" to "brews in 3 mins"). |
Recommended Rewrites
Passage: Home Office Recommendation
Before: Our compact brewer is the perfect choice for a small home office because it fits anywhere and makes great coffee quickly without any mess.
After: The [MODEL NAME] is optimized for home offices, measuring 12 inches tall with a 14oz reservoir. It utilizes a pressurized extraction system to brew a single cup in 160 seconds, eliminating the need for external filters or shared water lines. [SOURCE: USER MANUAL URL]
30-Day Priority List
- Immediate: Update
robots.txtto permit AI crawlers. - Week 1: Deploy Schema.org markup for the top 3 recommended machines.
- Week 2: Update the 'Home Office' and 'Small Spaces' sections using the provided specific-data rewrites.
Next steps
- Provide the URL for the product user manual to replace placeholders in the rewrites.
- Run the
scripts/ai_bot_robots_check.pyon your updated robots.txt to verify access. - Set a baseline in your prompt panel to track changes in 4 weeks.
Connects securely to your tools. The creator never sees your data.
About this skill
Get a clear, evidence-labeled plan for being cited and recommended by AI assistants.
When someone asks ChatGPT, Google AI Overviews or AI Mode, Claude or Perplexity for a recommendation, most sites never appear. This skill finds out why for your page and tells your agent exactly what to change.
What it does
- Finds the stage where you lose visibility: blocked, not retrieved, not selected, or not recommended. Fixes the right stage first instead of guessing.
- Checks crawler access by role. Separates search, user-fetch and training bots for OpenAI, Anthropic, Perplexity, Google and Bing, and flags accidental blocks. Bundled script parses your robots.txt.
- Scores the page on a 100-point rubric covering eligibility, extractability, on-page evidence, entity clarity, authority and freshness.
- Rewrites your key passages answer-first, with named subjects, sourced specifics and sub-question headings, shown as full before/after text.
- Plans honest off-page work: original data, expert commentary, genuine reviews, accurate listings, and disclosed community participation.
- Measures it: builds a 20 to 40 prompt panel, a run log, and a re-test schedule that accounts for the randomness of AI answers.
How it works
- Collects your page, topic, customer questions and competitors in one message.
- Builds the prompt panel and (if it can query engines) captures a baseline.
- Runs three audit tracks: crawl and eligibility, extractability and evidence, authority and corroboration.
- Ranks fixes by impact and effort into a 30/60/90 day plan.
- Delivers rewrites, an off-page plan, and a report with a "what could not be verified" section.
What makes it different
- Every recommendation carries an evidence tier. A: operator documentation or peer-reviewed research. B: third-party measurement. C: practitioner opinion. No myth presented as fact.
- Built on primary sources: Google Search Central, OpenAI and Anthropic crawler documentation, and the KDD 2024 GEO paper (Aggarwal et al.), with their limits stated.
- No shady tactics. It will not recommend hidden text, fake reviews, invented statistics or undisclosed sponsorship, and it uses placeholders instead of making up numbers.
- No guarantees. It says plainly that nobody can promise a citation, and it shows you how to measure what actually happened.
Where it stops
- It cannot see inside any AI engine and cannot guarantee ranking, citations, traffic or revenue.
- Engines change without notice. Facts are dated "last verified 2 November 2026" and the skill tells your agent to re-check crawler names and guidance when web access is available.
- It does not run JavaScript or check live indexing, and the helper scripts are heuristics.
- If your agent cannot query AI engines, the baseline is marked "pending" and you get the panel and log sheet to run yourself.
See it in action (700+ characters; use this, or replace with a real run of your own)
Example output, illustrating the report format on a fictional site.
Request: "Audit https://example-invoicing.com/late-payments for AI citation readiness. We want to be recommended for 'how to reduce late invoice payments' and 'best invoicing software for freelancers'. Competitors: FreshBooks, Wave."
Result (excerpt):
Top 3 findings
- (Tier A) robots.txt blocks
Claude-SearchBotandPerplexityBotfor the whole site through a blanket rule added for a staging copy. Both are search bots, so the page cannot appear in Claude or Perplexity search answers until this is fixed.GPTBot(training) is blocked separately and is the owner's choice. - (Tier C) The page opens with 90 words of preamble; the answer to "how to reduce late payments" starts in paragraph 4. 7 of 11 paragraphs begin with "It" or "This", so a lifted passage loses its subject.
- (Tier A research, B practice) The page makes five claims with no source. The GEO benchmark found that adding sourced statistics and quotations raised visibility, so each claim gets a
[STAT + SOURCE URL]slot. Score: 50/100 (A 8/20, B 12/25, C 7/20, D 6/10, E 9/15, F 8/10).
Rewrite, opening (before): "In today's fast-moving world, many freelancers wonder about invoicing..."
After: "Freelancers billing fewer than 20 clients a month cut late payments most by sending invoices within 24 hours of delivery, adding payment links, and scheduling automatic reminders. Below: the three habits, what each costs, and how to set them up. [Add sourced late-payment statistic + URL]"
30/60/90: day 1 robots fix; by day 30 rewrite three pages and add sourced figures; by day 60 pitch an original late-payment survey to two industry newsletters; re-test the 30-prompt panel at week 4 and monthly.
Not verified: Bing indexing status, CDN bot challenges (needs a live fetch), and baseline results for ChatGPT (no access in this run).
Limitations (optional field)
- Cannot guarantee citations or rankings.
- Needs web access to fetch pages and robots.txt; otherwise you paste them in.
- Heuristic scripts; no JavaScript rendering; no structured-data validation.
- Crawler names and engine behavior change; re-verify quarterly.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Marketing & SEO.
Limitations:
- Cannot guarantee citations or rankings.
88 - Needs web access to fetch pages and robots.txt; otherwise you paste them in.
89 - Heuristic scripts; no JavaScript rendering; no structured-data validation.
90 - Crawler names and engine behavior change; re-verify quarterly.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 9 days ago