Introduction

As large language models (LLMs) become widespread, search engine results pages (SERPs) are being reshaped by AI Overviews and intelligent summarization. For content producers and agent developers, how to enable AI models to accurately crawl, cite, or display specific information has become a new technical challenge. Traditional SEO logic needs to shift toward GEO (Generative Engine Optimization).

Plugin Overview

satan9394/dsh-geo-seo is a DeepSeek Harness (DSH) skill plugin focused on GEO-first AI search optimization. It aims to help developers evaluate how their content performs and how citable it is in mainstream AI search engines.

Core Features

The plugin provides the following core capabilities:

  1. Coverage of mainstream AI search engines: Supports crawling and optimization for ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews.
  2. Citability assessment: Provides a citability score to quantify the likelihood of content being directly cited by AI.
  3. Protocol support: Supports AI crawler and llms.txt protocol detection.
  4. Brand and structured data audit: Checks brand mentions and Schema E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) structured data.
  5. Parallel audit: Performs an in-depth audit using five parallel subagents.

Installation

The official documentation does not provide specific installation commands. Please visit the GitHub repository to obtain the source code and integrate it into your DSH environment.

Notes

  • Permission requirements: The plugin runs with the permissions of the current DSH process. Please ensure the environment is configured correctly.
  • License: The license is not explicitly specified. Be sure to review the source code before installation.
  • Inspiration: The plugin design was inspired by zubair-trabzada/geo-seo-claude (9.4k★).

Conclusion

This plugin provides agent developers with basic tools for monitoring and optimizing AI crawling behavior, serving as a starting point for building a content ecosystem for AI search. The complete source code and directory information can be found on GitHub.