sadassimov/geemu-skill

GEEMu 技能

🇨🇳 中文版

Google Earth Engine 主工具（GEEMu）。

Google Earth Engine Master Utility（GEEMu）是一个紧凑且完整的 OpenAI Codex 和 Claude Code 技能文件夹。它很有帮助 编写 Google Earth Engine （GEE） Python 工作流程，使用 geemap。其设计对象为 研究级遥感任务中，助理必须首先检查 了解研究设计，理解数据，以及 边界，然后编写可重复的GEE Python代码。

兼容性

• OpenAI Codex：将GEEMu安装为本地技能文件夹，然后请求Codex使用。 GEEMu用于GEE/geemap任务。
• Claude Code：安装同一个文件夹，位于 Claude 技能目录下，然后 让Claude Code使用GEEMu。
• 该技能基于文件。它不需要单独的向量搜索服务器， 嵌入模型下载、私有凭证或原始开发 工作空间。

安装

GitHub仓库：

https://github.com/sadassimov/geemu-skill

安装完整文件夹——不要跳过任何文件。 这项技能取决于 整个套餐，尤其是本地知识库 gee_vector_db/chunks.jsonl 和 gee_vector_db/documents.jsonl（数十 MB）：

GEEMu Skill

🇨🇳 中文版

Google Earth Engine Master Utility (GEEMu).

Google Earth Engine Master Utility (GEEMu) is a compact full skill folder for OpenAI Codex and Claude Code. It helps write Google Earth Engine (GEE) Python workflows with geemap. It is designed for research-grade remote sensing tasks where the assistant must first check the local environment, understand the study design, reason about data and boundaries, then write reproducible GEE Python code.

Compatibility

• OpenAI Codex: install GEEMu as a local skill folder, then ask Codex to use GEEMu for GEE/geemap tasks.
• Claude Code: install the same folder under the Claude skills directory, then ask Claude Code to use GEEMu.
• The skill is file-based. It does not require a separate vector-search server, embedding model download, private credentials, or the original development workspace.

Installation

GitHub repository:

https://github.com/sadassimov/geemu-skill

Install the complete folder — do not skip any files. The skill depends on the whole package, especially the local knowledge base gee_vector_db/chunks.jsonl and gee_vector_db/documents.jsonl (tens of MB): these JSONL files are what the retrieval workflow searches, so the skill will not work without them. git clone and GitHub’s “Download ZIP” already include every file — just don’t hand-pick files, exclude large/.jsonl files, or use a shallow/sparse/partial checkout.

Option A: Clone From GitHub

For OpenAI Codex on Windows PowerShell:

git clone https://github.com/sadassimov/geemu-skill.git "$env:USERPROFILE\.codex\skills\GEEMu"

For Claude Code on Windows PowerShell:

git clone https://github.com/sadassimov/geemu-skill.git "$env:USERPROFILE\.claude\skills\GEEMu"

For OpenAI Codex on macOS/Linux:

git clone https://github.com/sadassimov/geemu-skill.git ~/.codex/skills/GEEMu

For Claude Code on macOS/Linux:

git clone https://github.com/sadassimov/geemu-skill.git ~/.claude/skills/GEEMu

Option B: Download ZIP

1. Open https://github.com/sadassimov/geemu-skill.
2. Download the repository as a ZIP file.
3. Unzip it and rename the folder to GEEMu.
4. Put the folder here for Codex: ~/.codex/skills/GEEMu.
5. Put the folder here for Claude Code: ~/.claude/skills/GEEMu.

Windows users can use these equivalent locations:

• Codex: %USERPROFILE%\.codex\skills\GEEMu
• Claude Code: %USERPROFILE%\.claude\skills\GEEMu

Architecture

GEEMu uses a layered design:

• SKILL.md: the main instruction layer. It tells the assistant how to start every GEE task, including project ID, authentication, geemap import, proxy