Skills Manage: The Desktop App That Unifies AI Coding Agent Skills Across 27+ Platforms

Managing AI coding agent skills across multiple platforms is a growing challenge for developers. Each AI coding tool – Claude Code, Cursor, Windsurf, Copilot, Aider – stores its skills in a different directory, making it tedious to install, update, and organize skills consistently. Skills Manage is an open-source Tauri v2 desktop application that solves this problem by providing a single interface to manage skills across 27+ AI coding platforms, using symlinks to maintain one source of truth while deploying to every platform you use.

Skills Manage Architecture

What is Skills Manage?

Skills Manage (iamzhihuix/skills-manage) is a desktop application built with Tauri v2, React 19, and Rust that follows the Agent Skills open pattern. It uses ~/.agents/skills/ as the canonical central directory for all skills, then creates symlinks from that central location to each platform’s individual skills directory. This means you install a skill once, and it automatically becomes available across every supported platform.

The app provides a rich graphical interface for browsing, installing, organizing, and discovering skills – eliminating the need to manually copy SKILL.md files between directories or manage symlinks by hand.

Architecture Overview

The architecture diagram above illustrates the three-layer design of Skills Manage. Let’s break down each component:

User Interface Layer (React 19)

The frontend is built with React 19 and TypeScript, using shadcn/ui components for a polished, responsive interface. Key UI technologies include:

  • shadcn/ui provides consistent, accessible component patterns including dialogs, drawers, cards, and virtualized lists for handling large skill libraries efficiently
  • Zustand manages application state with lightweight, performant stores for skills, platforms, collections, marketplace data, and settings
  • react-i18next enables bilingual support (English and Chinese) with automatic language detection
  • react-router-dom v7 handles client-side routing across the six main views: Central Skills, Platform, Discover, Marketplace, Collections, and Settings
  • react-markdown renders SKILL.md content with full GFM (GitHub Flavored Markdown) support for rich skill previews

Tauri v2 IPC Bridge

The Tauri v2 runtime serves as the bridge between the React frontend and the Rust backend. It provides:

  • Secure IPC (Inter-Process Communication) for all skill operations
  • Native file system access for reading, writing, and creating symlinks
  • Platform-specific dialog APIs for file selection and confirmation
  • Shell access for executing system commands when needed

Rust Backend Layer

The Rust backend handles all heavy lifting, organized into six command modules:

  • Scanner discovers skills by walking platform directories and parsing SKILL.md frontmatter
  • Linker manages symlink creation and removal between the central directory and platform directories
  • Marketplace syncs skill metadata from official publishers (Microsoft, Anthropic, GitHub, etc.) with authenticated requests and retry fallback
  • Collections provides CRUD operations for organizing skills into named groups
  • Discover scans local project directories for project-level skill libraries
  • Settings manages configuration including AI provider keys, GitHub PAT, and custom platform directories

Storage Layer

All data is stored locally in SQLite (~/.skillsmanage/db.sqlite) using WAL mode for concurrent read/write performance. The database tracks skills, installations, collections, marketplace caches, and AI-generated explanations. No data is sent to external servers unless you explicitly use marketplace sync, GitHub import, or AI explanation features.

Supported Platforms

Skills Manage Platform Ecosystem

Understanding the Platform Ecosystem

The platform ecosystem diagram shows how Skills Manage connects a single central skills directory to 27+ individual AI coding platforms. The central hub at the top (~/.agents/skills/) is the single source of truth, and symlinks fan out to each platform’s directory.

Coding Platforms (21 platforms)

Skills Manage supports the following AI coding platforms, each with its own dedicated skills directory:

Platform Skills Directory Category
Claude Code ~/.claude/skills/ Coding
Codex CLI ~/.agents/skills/ Coding
Cursor ~/.cursor/skills/ Coding
Gemini CLI ~/.gemini/skills/ Coding
Trae ~/.trae/skills/ Coding
Factory Droid ~/.factory/skills/ Coding
Junie ~/.junie/skills/ Coding
Qwen ~/.qwen/skills/ Coding
Trae CN ~/.trae-cn/skills/ Coding
Windsurf ~/.windsurf/skills/ Coding
Qoder ~/.qoder/skills/ Coding
Augment ~/.augment/skills/ Coding
OpenCode ~/.opencode/skills/ Coding
KiloCode ~/.kilocode/skills/ Coding
OB1 ~/.ob1/skills/ Coding
Amp ~/.amp/skills/ Coding
Kiro ~/.kiro/skills/ Coding
CodeBuddy ~/.codebuddy/skills/ Coding
Hermes ~/.hermes/skills/ Coding
Copilot ~/.copilot/skills/ Coding
Aider ~/.aider/skills/ Coding

Lobster Platforms (5 platforms)

The Lobster ecosystem platforms are also supported:

Platform Skills Directory
OpenClaw ~/.openclaw/skills/
QClaw ~/.qclaw/skills/
EasyClaw ~/.easyclaw/skills/
AutoClaw ~/.openclaw-autoclaw/skills/
WorkBuddy ~/.workbuddy/skills-marketplace/skills/

Central Skills

The central directory (~/.agents/skills/) serves as the canonical location shared with Codex CLI, ensuring one source of truth for all skill content.

Marketplace Sources

The app connects to official skill publishers with over 2,500 skills available:

  • Microsoft (404 skills) - Azure skills
  • GitHub (331 skills) - Copilot skills
  • Anthropic (289 skills) - Claude Code skills across 9 repos
  • Sentry (244 skills) - Monitoring skills
  • Vercel Labs (214 skills) - Agent skills
  • Firecrawl (168 skills) - CLI skills
  • Plus 50+ more publishers including PostHog, Cloudflare, HashiCorp, Firebase, and more

Custom platforms can also be added through the Settings view, making Skills Manage extensible for any future AI coding tool.

Key Features

Central Skill Library

The Central Skills view shows all skills installed in ~/.agents/skills/. This is your single source of truth – every skill here can be symlinked to any supported platform with one click. The view supports:

  • Fast search with deferred queries and lazy indexing
  • Virtualized lists for smooth scrolling through large skill libraries
  • Markdown preview with syntax highlighting
  • Raw source view for inspecting SKILL.md content
  • AI-powered explanation generation using Claude, GLM, MiniMax, Kimi, DeepSeek, or OpenRouter

Per-Platform Install and Uninstall

Each platform has its own view showing which skills are currently installed. Skills Manage handles three types of installations:

  • Symlink (default) – creates a symbolic link from the platform directory to the central skill, so updates to the central copy are immediately reflected everywhere
  • Copy – duplicates the skill file into the platform directory for isolated use
  • Native – skills that already exist in the platform directory without any linking

Claude Code has special handling: it also surfaces marketplace plugin directories under ~/.claude/plugins/marketplaces/* as read-only entries, giving you visibility into plugin skills alongside native ones.

Collections

Collections let you organize skills into named groups for batch operations. For example, you could create a “Frontend Development” collection containing skills for React, Tailwind CSS, and TypeScript, then install the entire collection to multiple platforms at once.

Discover

The Discover feature scans your local disk for project-level skill libraries. If you have skills embedded in project directories (e.g., .claude/skills/ inside a project), Discover finds them and lets you install them centrally.

Marketplace Browsing and GitHub Import

The Marketplace view lets you browse skills from official publishers with authenticated GitHub requests and retry fallback. You can:

  • Browse by publisher (Microsoft, Anthropic, GitHub, etc.)
  • Preview skill content before installing
  • Download and install skills directly from the marketplace
  • Import skills from any GitHub repository using the Import Wizard

The GitHub Import Wizard supports authenticated requests, so you can import from private repositories using a GitHub Personal Access Token stored securely in the local SQLite database.

AI Explanation Generation

Skills Manage can generate AI-powered explanations of any skill using multiple providers:

Provider Model Regions
Claude claude-sonnet-4-20250514 International
Zhipu GLM glm-5 China, International
MiniMax MiniMax-M2.7 China, International
Kimi kimi-k2.5 China
DeepSeek DeepSeek-V3.2 China
OpenRouter anthropic/claude-sonnet-4.6 International
Custom User-configured All

Explanations are cached locally in the database, so you only pay for generation once per skill per language.

Skill Management Workflow

Skills Manage Workflow

Understanding the Skill Lifecycle

The workflow diagram above illustrates the complete lifecycle of a skill from discovery to activation across platforms. Let’s walk through each phase:

Phase 1: Discovery

Skills can enter your library through three channels:

  1. Local Scan – The Scanner module walks all configured skill directories and parses SKILL.md frontmatter to discover existing skills on your machine
  2. Marketplace Browse – Browse the official marketplace with 2,500+ skills from 50+ publishers, preview content, and download directly
  3. GitHub Import – Use the Import Wizard to pull skills from any GitHub repository, with authenticated access for private repos

Phase 2: Selection and Preview

Once you’ve found a skill, you can:

  • Preview the full SKILL.md content with Markdown rendering
  • Optionally generate an AI explanation to understand what the skill does before installing
  • Select individual skills or multiple skills for batch installation

Phase 3: Installation

The installation process follows the symlink pattern:

  1. The skill content is stored in the central directory (~/.agents/skills/skill-name/SKILL.md)
  2. Symlinks are created from each target platform’s directory pointing back to the central copy
  3. This ensures one source of truth – update the central copy, and all platforms see the change immediately

Phase 4: Organization

After installation, you can optionally organize skills into collections:

  • Create named collections like “DevOps”, “Frontend”, “Security”
  • Add skills to collections for logical grouping
  • Batch-install entire collections to new platforms with one action

Phase 5: Active Use

Once installed, skills are immediately active across all linked platforms. The next time you use Claude Code, Cursor, Windsurf, or any other supported tool, the skills are automatically available.

Installation

Prerequisites

Download Prebuilt

The easiest way to get started is to download the latest release:

# Download from GitHub Releases
# https://github.com/iamzhihuix/skills-manage/releases/latest

Currently, prebuilt packages are available for Apple Silicon macOS (.dmg and .app.zip). Other platforms can run from source.

Run from Source

# Clone the repository
git clone https://github.com/iamzhihuix/skills-manage.git
cd skills-manage

# Install frontend dependencies
pnpm install

# Run in development mode
pnpm tauri dev

The Vite dev server runs on port 24200.

macOS Unsigned Build

If macOS shows a warning about the app being damaged, run:

xattr -dr com.apple.quarantine "/Applications/skills-manage.app"

Then launch the app again from Finder.

Validation

# Run frontend tests
pnpm test

# Type checking
pnpm typecheck

# Linting
pnpm lint

# Rust backend tests
cd src-tauri && cargo test

# Rust linting
cd src-tauri && cargo clippy -- -D warnings

Usage Examples

Installing a Skill to Multiple Platforms

# Skills are stored centrally in ~/.agents/skills/
# When you install a skill through the UI, symlinks are created:
#
# ~/.agents/skills/frontend-design/SKILL.md  (central copy)
# ~/.claude/skills/frontend-design/SKILL.md  -> symlink to central
# ~/.cursor/skills/frontend-design/SKILL.md  -> symlink to central
# ~/.windsurf/skills/frontend-design/SKILL.md -> symlink to central

Browsing the Marketplace

The Marketplace view shows skills organized by publisher. You can filter by category tags:

Tag Description
Frontend React, Tailwind, UI development
Backend Server-side, APIs, databases
E-commerce Shopify, Stripe, payment integration
App Dev Flutter, Expo, mobile development
DevOps Terraform, Pulumi, infrastructure
AI/ML LangChain, Hugging Face, model integration
Database Prisma, Redis, data management
Security Auth0, Clerk, authentication
Testing Sentry, Datadog, monitoring
Docs/Design Notion, Figma, documentation

Creating a Collection

Collections help you organize skills for batch operations:

  1. Navigate to the Collections view
  2. Click “New Collection” and give it a name and description
  3. Use the Skill Picker to add skills from your library
  4. Install the entire collection to one or more platforms

Using AI Explanations

To generate an AI explanation for any skill:

  1. Open a skill’s detail view
  2. Click the “Explain” button
  3. Choose your preferred AI provider (Claude, GLM, MiniMax, etc.)
  4. The explanation is generated and cached locally for future reference

Tech Stack

Layer Technology
Desktop framework Tauri v2
Frontend React 19, TypeScript, Tailwind CSS 4
UI components shadcn/ui, Lucide icons
State management Zustand
Markdown react-markdown with remark-gfm
i18n react-i18next, i18next-browser-languagedetector
Theming Catppuccin 4-flavor palette
Backend Rust (serde, sqlx, chrono, uuid)
Database SQLite via sqlx (WAL mode)
Routing react-router-dom v7

Privacy and Security

Skills Manage is designed with privacy as a core principle:

  • Local-first storage – All metadata, collections, scan results, settings, and cached AI explanations stay in ~/.skillsmanage/db.sqlite or local skill directories
  • No telemetry – The app does not include analytics, crash reporting, or usage tracking
  • Feature-driven network access – Outbound requests only happen when you explicitly use marketplace sync, GitHub import, or AI explanation generation
  • Local credentials – GitHub PAT and AI API keys are stored in the local SQLite settings table (not encrypted at rest by the app)

Comparison: Skills Manage vs Manual Skill Management

Aspect Manual Management Skills Manage
Installation Copy SKILL.md to each directory One-click install with symlinks
Updates Manually update each copy Update central copy, all platforms sync
Discovery Browse GitHub repos manually Marketplace with 2,500+ skills
Organization No grouping system Named collections with batch install
Multi-platform Copy to 27+ directories Symlink to all platforms at once
Preview Open file in editor Markdown preview with AI explanation
Search Manual file system search Fast indexed search with virtualization
Project skills Unknown Discover scan finds project-level skills

Project Structure

skills-manage/
+-- src/                        # React frontend
|   +-- components/             # UI components
|   +-- i18n/                   # Locale files and i18n setup
|   +-- lib/                    # Frontend helpers
|   +-- pages/                  # Route views
|   +-- stores/                 # Zustand stores
|   +-- test/                   # Vitest + RTL tests
|   +-- data/                   # Official sources and AI providers
+-- src-tauri/                  # Rust backend
|   +-- src/
|       +-- commands/           # Tauri IPC handlers
|       +-- db.rs              # SQLite schema, migrations, queries
|       +-- lib.rs             # Tauri app setup
|       +-- main.rs            # Desktop entry point
+-- public/                     # Static assets
+-- CHANGELOG.md                # English changelog
+-- CHANGELOG.zh.md             # Chinese changelog
+-- release-notes/              # GitHub release notes

Conclusion

Skills Manage addresses a real and growing pain point in the AI coding ecosystem: as the number of AI coding agents multiplies, managing skills across all of them becomes increasingly complex. By centralizing skill storage with ~/.agents/skills/ and using symlinks to deploy to each platform, Skills Manage provides a clean, efficient solution that scales with the ecosystem.

The combination of a Tauri v2 desktop app, React 19 frontend, and Rust backend delivers native performance with a modern UI. Features like marketplace browsing, GitHub import, collections, discover scanning, and AI-powered explanations make it more than just a symlink manager – it is a comprehensive skill management platform.

With support for 27+ platforms, 2,500+ marketplace skills from 50+ publishers, and a privacy-first local storage approach, Skills Manage is an essential tool for any developer working with multiple AI coding agents.

Watch PyShine on YouTube

Contents