ssw-plugin

SSW Plugin for Claude Code

Sun and Space Weather (SSW) toolkit for Claude Code. Download, preprocess, visualize, and apply machine learning to solar observation data from SDO, STEREO, and Solar Orbiter missions.

---

Table of Contents

• What is a Claude Code Plugin?
• Installation
  • Step 1: Add Marketplace
  • Step 2: Install Plugin
  • Step 3: Verify Installation
  • Step 4: Install Python Dependencies
• Plugin Management
• Skills Overview
• Workflow
• Skill 1: ssw-download
• Skill 2: ssw-prep
• Skill 3: ssw-ml
• Skill 4: ssw-viz
• Natural Language Usage
• Example Gallery
  • Example 1: Download STEREO Data
  • Example 2: Download Solar Orbiter Data
  • Example 3: Visualize STEREO Multi-Wavelength Pair
  • Example 4: Visualize Solar Orbiter Multi-Wavelength Pair
  • Example 5: Pixel Intensity Distribution
• For Plugin Developers
  • Plugin Structure
  • How to Create Your Own Plugin
  • Publishing as a Marketplace
• Dependencies
• License

---

What is a Claude Code Plugin?

Claude Code plugins are extensions that add new capabilities to the Claude Code CLI. A plugin can provide:

• Skills - Slash commands (e.g., /ssw-plugin:ssw-download) that give Claude specialized knowledge and workflows
• Agents - Custom subagent definitions for specialized tasks
• Hooks - Event handlers that respond to Claude Code lifecycle events
• MCP Servers - External tool integrations via Model Context Protocol
• LSP Servers - Code intelligence via Language Server Protocol

Plugins are distributed through marketplaces - Git repositories that catalog one or more plugins. Users add a marketplace, then install individual plugins from it.

---

Installation

Step 1: Add Marketplace

Open Claude Code and register the ssw-plugin marketplace:

Option A: Inside Claude Code (interactive)

/plugin marketplace add https://github.com/tykimos/ssw-plugin.git

Option B: From terminal (CLI)

claude plugin marketplace add https://github.com/tykimos/ssw-plugin.git

This clones the repository to ~/.claude/plugins/marketplaces/ssw-plugin/ and registers it in ~/.claude/plugins/known_marketplaces.json.

Step 2: Install Plugin

Option A: Inside Claude Code (interactive)

/plugin install ssw-plugin@ssw-plugin

The format is <plugin-name>@<marketplace-name>.

Option B: From terminal (CLI)

claude plugin install ssw-plugin@ssw-plugin

Option C: Interactive plugin manager

/plugin

This opens the plugin manager UI. Navigate to the Discover tab to browse and install available plugins.

Installation Scopes

You can control where the plugin is available:

# Available in all your projects (default)
claude plugin install ssw-plugin@ssw-plugin --scope user

# Available only in current project, shared with team via git
claude plugin install ssw-plugin@ssw-plugin --scope project

# Available only in current project, not shared (gitignored)
claude plugin install ssw-plugin@ssw-plugin --scope local

Scope	Settings File	Shared via Git	Use Case
user	~/.claude/settings.json	No	Personal use across all projects
project	.claude/settings.json	Yes	Team-wide plugin for a project
local	.claude/settings.local.json	No	Personal use for one project

Step 3: Verify Installation

After installation, verify the plugin is active:

/plugin

You should see ssw-plugin listed as Enabled. The following slash commands should be available:

• /ssw-plugin:ssw-download
• /ssw-plugin:ssw-prep
• /ssw-plugin:ssw-ml
• /ssw-plugin:ssw-viz

You can also test by typing /ssw- and checking autocomplete suggestions.

Step 4: Install Python Dependencies

The plugin skills require Python packages. Install them in your environment:

# Core: solar data download and preprocessing
pip install git+https://github.com/sswlab/ssw-tools
pip install sunpy matplotlib astropy aiapy

# Optional: for ML tasks (ssw-ml skill)
pip install torch torchvision scikit-image

# Optional: for logging in batch processing
pip install loguru

---

Plugin Management