Websites & apps · SKILL

GEO: Generative Engine Optimization

Check AI search mentions, citations and recommendations, then test changes with a repeatable query panel and verifiable sources.

Skill introduction

For website and brand content owners assessing an entity’s visibility in AI search. Use current pages, reliable sources, and a fixed question set for each language and market to deliver evidence-backed gaps, priorities and a repeatable baseline.

How to verify the work

Use the same question set to open target pages, answers and citations. Check the entity, cited sources and crawl access. Record dates and the number of valid responses. On reruns, check accurate mentions, official-site citations, recommendations and visits separately.

When to hand it over

Hand supported content or technical changes to the content or site owner once the evidence and baseline are in place. Flag untraceable sources or changed sample conditions, and hold the related attribution while gathering evidence.

What to measure

Rerun the same questions to log changes in accurate mentions, official-site citations and recommendations. Track visits and platform variability separately.

What it does

Use a fixed question set to observe accurate mentions, citations and recommendations in AI search.

What it helps with

Use real sources and repeat measurements to find specific gaps in content, crawl access and citations.

Why it matters

Keep comparable evidence for long-term content decisions and distinguish platform variability from the effects of changes.

PDF · Chinese

Guide

Core guidance, examples and a revision worksheet. PDF in Chinese.

V1.2 · 21 pages · 322 KB

Usage guide

For websites or brands with a defined AI search goal. Review content evidence, technical access and third-party citations. Separate observations, inferences and open questions, and track accurate mentions, official-site citations, recommendations and traffic as distinct results.

What to prepare

  • The website, entity, audience, language and market to study
  • Current pages, verifiable sources and relevant AI answers
  • Available traffic, citation or manually annotated observation data

What you receive

  • Evidence-backed content and technical findings
  • Prioritized rewrites or content briefs
  • A repeatable question panel and separately measured outcomes

Who it is for

  • Investigating inaccurate or missing mentions in AI search
  • Giving editors evidence-backed content tasks
  • Training a team to distinguish mentions, citations and recommendations

How the work progresses

  1. 01

    Define the questions and target entity

  2. 02

    Define a repeatable panel for the entity, language and market, and record the date and baseline denominator.

  3. 03

    Inspect pages, answers and citation evidence

  4. 04

    Identify the specific content or access gap

  5. 05

    Make changes and repeat the same observations

  6. 06

    Compare mentions, citations, recommendations and traffic separately, and report uncertainty.

Getting started

Download and extract the complete ZIP. Read the README if included, or start with SKILL.md. Install the complete folder containing SKILL.md in your AI tool’s skills directory under the skill name. In a multi-version repository, choose the version you need and retain its supporting files. Try a small task with the required inputs and review the result.

Scope and limits

  • The public package contains complete Chinese and English Skills. Choose one to avoid duplicate activation.
  • The download is a complete repository ZIP pinned to a commit. Install a Skill from skills/; it is not an automatic installer.
  • The offline tool aggregates annotated data and does not verify facts. Inclusion, rankings, traffic and business results are not guaranteed.
  • The Chinese manual is V1.2. The pinned public Chinese and English source is 1.0.0 under the MIT license.

Updates for this Skill

Back to My Skills All updates
SEARCH