SkillFoxx

HeiGe-GEO-SEO skill

A GEO and SEO skill for the Chinese market: page scoring, rewrites for AI citation, llms.txt and competitor comparison

  • 90 ★ on the repository
  • Last commit
Skill

Minimal access

Why this level

  • Scripts work offline on the standard library, read pages you point to and write report files into the project folder
All reasons and checks

heigeai/heige-geo-seoAuthor:HeiGeAi

Install

Manual install

Text for your agent

Copy `adapters/claude-code/heige-geo-seo/` from HeiGeAi/HeiGe-GEO-SEO into `~/.claude/skills/`. Run scripts as `python3 scripts/geo_cli.py <command>` from the skill folder and show paths before writing files.

Access and checks

How we review

Minimal accessThe tool runs nothing and makes no network calls.

  • Scripts work offline on the standard library, read pages you point to and write report files into the project folder

This is third-party code. Review the repository files before installing.

What it does

A skill for optimizing content for AI answers with a China-first focus: Doubao, Qwen, DeepSeek, ERNIE, Tencent Yuanbao, plus ChatGPT, Perplexity, Claude and Gemini. The author argues that Chinese models have no public crawlers, citation builds on indexes and platform ecosystems, and `robots.txt` barely works. The knowledge library covers model classification, content-fetching logic, publishing channels, page architecture and rewrite methods. The `geo_cli.py` script gives 29 commands: a page score on 6 dimensions and 22 items, `rewrite` builds instructions for any LLM, `brief` builds a content brief, `sov` computes brand share of voice, `diagnose` names seven reasons a page is not cited, `files` and `schema` create `llms.txt`, `robots.txt` and JSON-LD, `compare` compares with a competitor and `playbook` outputs a printable HTML plan. It runs on the Python standard library.

Who it is for. SEO specialists and content teams promoting sites and brands in Chinese AI search and in ChatGPT and Perplexity.

Good fit when

  • You need to score a page on 22 items and learn why it is not cited (`diagnose`)
  • You want rewrite instructions for Perplexity or Doubao with the `rewrite` command
  • A site needs `llms.txt`, `robots.txt` patches and JSON-LD markup in one command
  • You need to compare your page with a competitor's and see element gaps

Not a fit when

  • Promotion targets the Russian market and Yandex only: the skill is built for Chinese models and search engines
  • You need automatic measurement of brand mentions in AI: the skill computes metrics from records you collect yourself
  • You need a platform with an interface and monitoring: this is a set of scripts and knowledge

Example request

Score this page on the 22 GEO items and write rewrite instructions for Doubao and Perplexity

Limitations

  • The main materials, the README and descriptions are in Chinese, with an English section at the end of the README.
  • Scripts work offline on the Python standard library: collecting AI answers goes through your agent and not a built-in API.
  • Audience figures for the models (such as Doubao user numbers) were taken by the author at writing time and age.
  • Citation growth advice rests on public research and the author's experience, and the skill gives no guarantee of results in AI search.

How to disable. Delete the `heige-geo-seo` skill folder from `~/.claude/skills/` or your agent's directory.

FAQ

How do I install the skill in Claude Code?

Copy the ready package from `adapters/claude-code/heige-geo-seo/` into `~/.claude/skills/` and it triggers on GEO and AI search requests. Codex, OpenClaw and Hermes have their own packages in `adapters`.

How does Chinese GEO differ from the foreign one?

In the author's view Chinese AI has no public crawlers and relies on search indexes and platform ecosystems, so the goal is indexing and holding a platform, and `robots.txt` barely matters.

Which commands does `geo_cli.py` have?

Page scoring, `rewrite`, `brief`, `prompts`, `sov`, `diagnose`, `files`, `schema`, `report` in SARIF, `batch`, `attribution`, `agentready`, `compare` and `playbook`. 29 commands in total.

Are external dependencies needed?

No, the scripts use only the Python standard library and work offline. The author reports 447 passing tests.

What does the `playbook` command give?

A full plan: bottleneck location across eight layers, diagnosis, source strategy, paragraph rewrite guidance, measurement collection and an action list. It outputs printable HTML.

Discussion

No questions or reviews yet.

Skills

24 skills for GEO: page audits for AI answers, articles and comparisons, brand mention sampling in ChatGPT and DeepSeek, Word and PDF reports

SkillBroad access917

Three GEO skills: a one-off page audit from a URL, step-by-step optimization and a platform for long-term monitoring.

SkillExtended access238

A CLI, Python library and MCP server that scores how ready a site is to be cited by ChatGPT, Perplexity and AI Overviews

CLIExtended access1.3K

SEO and AI-search optimization skill: a 100-point site score, an action plan, ready fixes and 20 diagnostic scripts.

SkillExtended access103

A skill and CLI for GEO: check whether AI models can read your site, find who they recommend instead, draft fixes and re-measure.

PluginExtended access607

A Claude Code skill set for auditing sites for AI search: citability, AI crawler access, llms.txt, schema and reports

SkillExtended access11.6K

Comparison with similar tools

Cards that share the main search query, and how they differ.

ToolOfficialRepository starsMain difference
HeiGe-GEO-SEO (this card)no90
GEO-SEO Claudeno11,640a site audit for AI search aimed mainly at ChatGPT and Perplexity, while this skill covers Chinese models and gives rewrite instructions
Yao GEO Skillsno91724 GEO skills with Word and PDF reports, while this set is one skill with a 29-command CLI and a scorecard

README in short

The README explains why the authors split Chinese GEO from the foreign one and lists the knowledge libraries. It then walks the versions from diagnosis to production and measurement layers, `geo_cli.py` commands with examples, the supported runtimes (Claude Code, Codex, OpenClaw, Hermes) and the project structure. An English section ends it.

README badge

Original SKILL.md

---
name: heige-geo-seo
version: 1.11.3
description: >-
  HeiGe-GEO-SEO 是一套面向中国市场优先的 GEO(生成式引擎优化)+ SEO 内容优化系统。当用户需要让内容/网页被
  AI 大模型(豆包、通义千问、DeepSeek、文心、腾讯元宝、ChatGPT、Perplexity、Claude、Gemini)抓取并引用、
  做 GEO 优化、做 AI 搜索优化、优化网页被 AI 引用、写能被 AI 引用的内容、配 robots.txt 给 AI 爬虫、生成 llms.txt、
  生成 schema 结构化数据、规划内容发布渠道矩阵、给网页架构 SEO/GEO 建议时使用。
  Trigger in English on: GEO, generative engine optimization, AEO, answer engine optimization, AI search
  optimization, "get cited by ChatGPT/Perplexity/AI", optimize content for AI engines, llms.txt, AI crawler
  robots.txt, schema for AI citation. 触发词:GEO优化、生成式引擎优化、AI搜索优化、让AI引用、被AI引用、AEO、
  让豆包/千问/DeepSeek引用我、GEO SEO、AI可见度、内容被大模型抓取。
---

# HeiGe-GEO-SEO · 中国优先的 GEO + SEO 优化系统

> 集 GEO 学术方法论(KDD 2024 / ICLR 2026 两篇论文)+ 工程化评分框架 + 中国大模型抓取逻辑于一体。
> 别的 GEO 工具只盯 ChatGPT/Perplexity,这套**先打豆包、千问、DeepSeek、文心、元宝**,海外作对照。

## 这套工具解决什么

让你的内容和网页被 AI 大模型抓取并引用。覆盖四件事:**给谁优化**(大模型分类)、**怎么被抓到**(抓取逻辑)、**发哪里曝光**(渠道矩阵)、**网页怎么建**(架构建议)。

## 核心原则(先记这几条)

1. **国内 GEO 是求收录 + 占平台游戏,海外 GEO 是配 robots.txt 游戏。** 两套打法别混用,见 `knowledge/02-crawler-logic.md`。
2. **按排名分流:** 排名靠后的页面激进上引用类方法(可冲 +100%),排名第一的页面慎用(会 -30%),见 `methodology/geo-methods.md`。
3. **合规红线:引语/统计/引用只用真实素材,永不编造。** Keyword Stuffing 永久禁用(它是负效果)。改写后跑 GEU 护栏。
4. **人不可被绕过:** Brief 的专家洞察槽位和质量门的 BLOCK 裁决两个 checkpoint 不能自动跳过。GEO 惩罚没有真实洞察的水文。

## 怎么用(按需选路径)

### 完整客户交付 → 走十阶段工作流

读 `workflow/stages.md`,从 01 Intake 到 10 Re-audit 全程。两个人工 checkpoint(04 专家洞察、07 质量门 BLOCK)不可绕过。

### 只优化一篇已有内容 → 轻量路径

03 基线打分 → 05 改写 → 07 质量门复评。

1. 用 `scripts/geo_score.py` 给原内容/页面打基线分(6 维 22 项)
2. 按 `methodology/geo-methods.md` 选改写方法(看内容域 + 搜索排名档位)
3. 改写,默认组合"流畅度 + 加统计",需要时叠"标注来源"
4. 重打分对比,跑 GEU 护栏(KPC 矛盾度不超阈值)
5. 出 before/after 评分报告

Report a problem