yaojingang/yao-geo-skills

Yao GEO Skills – reusable “skill” packages for Generative‑Engine‑Optimization

What it is

• An open‑source library that bundles complete, version‑controlled skills (small, self‑contained AI‑agent workflows) for the whole GEO (Generative Engine Optimization) pipeline – from strategic audits to content production, page design, knowledge‑graph building and effect‑measurement.
• Each skill lives in its own folder and ships a SKILL.md (the entry‑point that defines the agent’s inputs, outputs and execution boundaries), a manifest.json, template briefs, and evaluation data that can be automatically validated with the provided Python script.

Why it matters

• GEO is the author’s term for the set of practices that make generated content discoverable and trustworthy on AI‑search platforms (DeepSeek, Doubao, Kimi, etc.).
• The repo moves beyond ad‑hoc prompts: it codifies repeatable processes, enforces quality checks, and produces ready‑to‑publish artefacts (HTML, DOCX, PDF, Markdown) that can be handed off to product, growth or content teams.

Key categories & example skills

Category	Sample skill	What it does
operations	yao-geoflow-cli	Drives bulk actions in a GEOFlow 2.1+ system (task export, knowledge‑base drafts, channel sync, Docker pre‑flight checks).
strategy	yao-geo-panorama-audit	Generates a brand‑wide GEO baseline, analysing AI‑answer visibility, competitor gaps and content gaps, then outputs a priority map.
page‑technical	yao-geo-page-audit	Crawls a site, checks crawlability, schema, AI‑extractability and produces a fix‑list plus multi‑format reports.
content‑production	yao-geo-title-optimizer	Produces a library of AI‑friendly titles with scoring, brand isolation and compliance filters.
knowledge‑assets	yao-geo-brand-graph	Turns corporate information into an entity‑relationship graph (Mermaid, JSON‑LD, triples) with trust scores.
measurement	yao-geo-effect-monitor	Designs a monitoring dashboard that tracks AI‑answer citations, attribution and monthly health reports.

How to use it

1. Clone the repo

   git clone https://github.com/yaojingang/yao-geo-skills.git
   cd yao-geo-skills
   

2. Validate the structure (optional but recommended):

   python3 scripts/validate_repository.py
   

3. Pick a skill – either checkout the whole skills/ folder or use sparse checkout to fetch only the one you need.
4. Run the skill – follow the SKILL.md in the skill’s directory; it describes the required input brief (found in templates/brief-template.md) and the command‑line or API call that triggers the agent.
5. Inspect the output – artefacts are generated in the formats listed in the skill’s documentation (HTML, DOCX, PDF, Markdown) and can be directly handed to designers, editors or product owners.

What makes it “open‑source‑ready”

• Reusable – each skill has a clear boundary and can be invoked in any similar project.
• Verifiable – evals/trigger_cases.json and evals/expected_artifacts.json let CI pipelines automatically check that a skill still produces the expected results.
• Privacy‑first – the repository includes a checklist that strips private data and internal API keys before publishing.
• Rich documentation – human‑friendly pages live under docs/skills/ and an interactive navigation page (index.html) lets users browse the catalog.

Who should look at this

• Teams that already run a GEOFlow system and need standardized automation for operations, content creation, or measurement.
• Content marketers, SEO specialists, or product growth engineers who want a plug‑and‑play way to make their pages AI‑search friendly.
• Anyone building AI‑search‑aware knowledge bases and needing a reproducible, audited workflow.

What it is not

• A simple list of prompts.
• A private, internal tool that depends on proprietary APIs.
• A generic “AI” library unrelated to generative‑engine optimisation.

---

All details above are taken directly from the repository’s README; no additional features have been inferred.