By ViryaZheng
Run a complete GEO (Generative Engine Optimization) pipeline from client intake through visibility audit, content gap analysis, brief creation, draft production, distribution, and re-audit, with interactive HTML review forms for non-technical stakeholders
Build the GEO brand context for a client by ingesting provided materials (PDF/DOCX/PPTX/XLSX, URLs, raw notes) and conducting structured AI-perception + competitor + community research. Writes brand_context.json validated against schemas/brand_context.schema.json. Use when starting a new client (clients/<slug>/inputs/ has materials but no brand_context.json), or when brand_context.json is marked status=stale. Downstream consumers: 02-audit, 03-gap, 04-content-brief, 05-production.
Run a Claude-native visibility audit. For each query in brand_context.target_queries, ask Claude to answer the query as a real user would (using WebSearch + WebFetch), then record whether the client brand is mentioned, at what position, and which URLs were cited. Outputs visibility_baseline.json validated against schema. This is the BASELINE — the reference point for 07-reaudit. Use after 01-intake; rerun for re-audit rounds.
Translate the visibility baseline + brand context into a ranked list of content actions that should move the needle. Identifies absent queries, contested queries, and competitor-displacement opportunities. Outputs content_priorities.json validated against schema. Use after 02-audit; consumed by 04-content-brief.
For one priority from content_priorities.json, generate a content brief with explicit slots for the human expert (founder/domain specialist) to fill with real cases, real data, and real opinions. This is the human-in-loop checkpoint — the moat that prevents the entire workflow from collapsing into "AI writing more AI-readable AI slop." Outputs briefs/<priority-id>.md. Use after 03-gap; consumed by 05-production ONLY after the expert has filled the marked slots.
Convert an expert-filled brief into a publishable draft. Applies Princeton KDD 2024 GEO techniques (statistics, quotations, citations, authoritative language) to maximize AI citation likelihood. Refuses to run on briefs that haven't been filled by the expert. Outputs drafts/<id>.md. Use after 04-content-brief produces a brief with status=ready-for-production.
Use when auditing content quality, E-E-A-T, publish readiness, or 内容质量/EEAT评分. Runs 80-item CORE-EEAT scoring with veto checks and fix plan.
Write SEO-optimized blog posts, landing pages, and long-form page copy following Google's E-E-A-T and Helpful Content guidelines. Handles new content creation from a keyword, topic, or brief, and full-page rewrites. Use when asked to "write a blog post", "create a landing page", "write content about X", "content for keyword X", "draft an article", "blog post about", "landing page for", "service page", "product page copy", "rewrite this page", "how-to guide", or "listicle". NOT for analyzing/auditing existing content (seo-geo-optimizer, content-quality-auditor), title tags or meta descriptions (meta-tags-optimizer), or keyword expansion (keyword-research) — those skills own their triggers.
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.
Entry point + orchestrator for the recomby-geo GEO (Generative Engine Optimization) workflow on OpenAI Codex CLI. Use when the user wants to run any stage of the GEO pipeline on a client folder — intake, visibility audit, content-gap analysis, content brief, draft production, distribution, or monthly re-audit — or asks to "run GEO", "audit AI search visibility", or "GEO this client". Codex has no bare slash commands, so this skill is how the 7 stages (that Claude Code runs as /01-intake … /07-reaudit) are driven on Codex. It routes to the per-stage specs in this plugin's commands/ and enforces the orchestration rules. Does not auto-fill expert content — the human-in-loop brief checkpoint is the moat.
Render an interactive, self-contained HTML companion for a GEO content brief (04-content-brief) or a publish-ready draft (05-production), so a NON-technical client reviewer (founder, organizer staff, the domain expert filling slots) can fill REQUIRED-FILL slots, leave section-level comments, and approve/return work in the browser instead of editing Markdown. Use when a brief or draft needs to go to a client/expert for review, or when building the briefs/index.html entry page for a client folder. The reviewer's input comes back as a JSON file that 04-content-brief Step 9 ingests. Visual quality is delegated to the frontend-design skill.
npx claudepluginhub viryazheng/recomby-geo --plugin recomby-geoBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
把 Agent、办公 CLI、GEO 专业知识 skills 整合成一套协作型 AI 员工方案——给你的 agent 装上 GEO skills,让它跟你的业务专家一起干 GEO 协作活儿。
不是「AI 替代员工」,是一个虚拟岗位的组装产物。四块拼起来才是 AI 员工,缺一不可:
业务专家 Agent
(你/客户团队,懂业务) (CC / Codex / ...,可接本地模型)
│ │
└──────────── ╳ ─────────┘
│
┌─────────────┴─────────────┐
│ │
办公 CLI 层 GEO Skills
(飞书/钉钉/Slack/Notion…) (← 本项目的主战场)
数据源 + 操作脚手架 专业知识赋能 agent
| 角色 | 提供者 | 本项目集合 / 开发 |
|---|---|---|
| 业务专家 | 你自己 / 客户团队 | — |
| Agent | CC / Codex / OpenCode / ... 生态 | ✓ 集合 → agents.md |
| 数据 + CLI 脚手架 | 飞书 / 钉钉 / Slack / Notion ... | ✓ 集合 → clis.md |
| GEO Skills | 本项目自研 + vendor 开源 | ✓ 主战场 → plugins/recomby-geo/skills/ |
我们做三件事:集合主流 Agent CLI、集合国内外办公软件 CLI、封装 GEO 专业知识 skills。
plugins/recomby-geo/(commands + skills + schemas)是本项目长期维护、未来持续优化的主要方向,MIT 许可证,可商用、可修改、可分发。
我们做的事是把 GEO 领域的专业知识封装成 agent 可调用的 skills,加上完整的 7 阶段协作工作流命令——这是仓库里我们持续投入、迭代、维护的核心资产。其中很多结构和方法论参考了 Princeton KDD 2024、Auriti Labs、awesome-geo 等开源工作,我们做的是组装、打磨、面向 AI 员工场景重新封装。
agents.md 和 clis.md 不是产品的另一部分——它们是方案架构说明页:讲清楚我们的 GEO skills 在「业务专家 + Agent + 办公 CLI + Skills」这套 AI 员工方案里处于什么位置,配合哪些 agent / CLI 工作。读者可以从那两页知道生态全貌,但主战场是 skills。
这套方案天然 Local-first——这是相比 SurferSEO / Frase / Clearscope 这类必须把客户数据上传到云端的 SaaS GEO 工具,最关键的差异化:
plugins/recomby-geo/skills/ 下封装了 6 个 GEO 专业知识 skill,模型根据 description 自动匹配调用——你不直接调用它们:
| Skill | 来源 | 用途 |
|---|---|---|
seo-geo-optimizer | 199-biotechnologies (MIT) | 主力,13 个 Python 脚本 |
content-writer | toprank (MIT) | 内容生产 |
content-quality-auditor | aaron-he-zhu (Apache 2.0) | E-E-A-T / 引用度审计 |
internal-linking-optimizer | aaron-he-zhu (Apache 2.0) | 内链优化 |
keyword-research | toprank (MIT) | 关键词扩展 |
meta-tags-optimizer | toprank (MIT) | 标题 / Meta |
加上 plugins/recomby-geo/commands/ 下 7 个我们自研的 GEO 工作流命令(/01-intake 到 /07-reaudit),共同构成完整的 GEO AI 员工能力包。这就是我们对外开源的核心产物。
许可证详见 LICENSE 与 THIRD_PARTY_LICENSES.md。
GEO Skills 装在哪个 agent、跟哪些办公 CLI 配合——这两份是方案架构说明,不是产品:
agents.md — 主流可加载 skills 的 AI agent CLI:CC、Codex CLI、OpenCode、Cline、Goose、字节 Trae 等clis.md — 让 agent 接客户业务数据:飞书 / 钉钉 / 企微 / 语雀(国内)+ Slack / Notion / GitHub / Linear / Google Workspace / Jira(国外)两份让你看清整套 AI 员工方案的全貌(agent 是身体、办公 CLI 是手脚),但核心要装的是 GEO Skills 本身。
GEO 天然不能全自动化——这是优势:
带来:客户决策方一听就懂、落地阻力小(业务专家不会有被替代恐惧)、价值归属清晰。
GEO 是协作型 AI 员工范式的最佳演示场。
# 1. 在 Claude Code 中安装本插件(GEO Skills + 7 阶段工作流命令)
/plugin marketplace add ViryaZheng/recomby-geo
/plugin install recomby-geo
# 2. 准备你的项目目录
mkdir -p clients/<your-project>/inputs
# 把业务资料(PDF / URL / notes)放进 inputs/
# 3. 启动协作工作流
/01-intake clients/<your-project>
# 按提示依次:/02-audit → /03-gap → /04-content-brief → ...
完整运行流程图:plugins/recomby-geo/orchestrator/run.md
不用 Claude Code? —— 在其他支持 skill/plugin 加载的 agent(见
agents.md)里手动拷贝plugins/recomby-geo/skills/即可加载。
完整 GEO 工作流 7 个阶段,重点是中段的「专家填写 brief」——agent 准备好框架,业务洞察必须由你填进去。这是 AI 员工协作本质的体现:
/01-intake → /02-audit → /03-gap → /04-content-brief
│
┌─────────┘
▼
[业务专家填写洞察槽位]
│
▼
/05-production → /06-distribution → 发布 → 7 天后 /07-reaudit
各阶段命令文档:plugins/recomby-geo/commands/
关键约束:
/05-production 在 brief 状态不是 ready-for-production 时硬拒绝——专家槽位没填完,agent 不会自动用 AI 内容代填。这是协作的硬门槛。/02-audit 用无客户上下文的子 agent 跑可见性基线,模拟陌生用户对 AI 引擎的真实提问。20 SEO/GEO skills and 5 commands on one shared contract for keyword research, content creation, technical audits, schema markup, monitoring, quality gates, entity truth, and campaign memory.
SEO & GEO content toolkit: /geo-lint validates content with 92 deterministic rules, /content-creator builds a self-configuring content pipeline with auto-discovery, brand voice, keyword research, and lint-fix validation until zero violations.
Scored SEO audits, technical health checks, Schema.org JSON-LD, E-E-A-T content scoring, domain authority rating, and GEO for AI Overviews, ChatGPT, and Perplexity. 24 diagnostic scripts. Modes: Audit → Plan → Execute.
Free SEO + AEO + GEO toolkit by Akii. All skills run on your Claude Code session model (Claude Sonnet / Opus / etc). Aligned with Google's AI Optimization Guide for Google AI Overviews + AI Mode. Audit sites, plan content, optimize pages, generate schema, cluster keywords. AI Visibility Score specifically calls the Akii backend (Llama 4 / DeepSeek as LLM judge against public footprint) + per-engine public-signal proxy — not direct queries of ChatGPT / Claude / Gemini / Perplexity / Copilot (those are an Akii platform feature). Auto-detects Ahrefs, GSC, Apify MCPs for richer data. Sends anonymized telemetry (plugin version + hashed machine UUID + OS, once per 24h) to akii.com — see PRIVACY.md, opt out via AKII_PLUGIN_DISABLE_TELEMETRY=1.
AI-visibility engineering, the open way — the THL GEO / AI-search audit skill suite (citability, AI-crawler access, llms.txt, schema, brand mentions, technical, content, reporting) plus the THL-original agent-readiness-scan, organised around the THL GEO Method.
GEO 引用追踪 + 优化。每周跑 Claude/GPT/Perplexity/Gemini 固定查询检测是否被引用,对未被引用的页补 Citable Statistics + llms.txt v2。