Skip to content

aussiegingersnap/cursor-skills

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

37 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Cursor Skills

A curated collection of skills for Cursor AI. Skills give your AI agent specialized capabilities - from building UIs to managing databases to creating documents.

Installation

Option 1: Cursor Settings (Recommended - Cursor 2.4+)

The simplest way to use skills. No server setup required.

  1. Open Cursor Settings (Cmd+Shift+J)
  2. Go to the Rules tab
  3. Click Add RuleRemote Rule (GitHub)
  4. Enter: https://github.com/aussiegingersnap/cursor-skills
  5. Select the skills you want to import

Skills are copied to your .cursor/skills/ directory and automatically discovered.

Option 2: Manual Copy

Clone and copy specific skills to your project:

git clone https://github.com/aussiegingersnap/cursor-skills /tmp/cursor-skills
cp -r /tmp/cursor-skills/skills/ui-design-system .cursor/skills/

Option 3: MCP Server (Advanced)

For global skill access across all projects or legacy Cursor versions.

  1. Copy .cursor/mcp_example.json to ~/.cursor/mcp.json (global) or .cursor/mcp.json (project)
  2. Update the path to your cloned mcp/skills_mcp.py
  3. Reload Cursor
  4. Copy main_rule.mdc contents to Cursor Settings → Rules (for global setup)

See MCP Setup below for details.


Comparison: Native vs MCP

Feature Native (Cursor 2.4+) MCP Server
Setup None - auto-discovered Python + uv required
Scope Project-specific Global or project
Discovery Automatic Via list_skills tool
Import Cursor Settings UI Via import_skill tool
Best for Most users Advanced setups, global skills

Recommendation: Use native skills for most projects. Use MCP only if you need global skills across all projects.


Available Skills

Core Skills (Recommended for all projects)

Skill Description
feature-build Complete feature development lifecycle with phases
skill-creator Create new skills with guided workflow
documentation Project documentation standards
versioning Semantic versioning with CHANGELOG
judge Quality review for complex changes

UI & Frontend (ui-, state-)

Skill Description
ui-design-system Linear/Notion-inspired UI patterns with style guide
ui-principles Minimal, crafted UI principles
nextjs-16 Next.js 16 App Router patterns
state-effector Effector reactive state management
state-tanstack Tanstack Query + Zustand patterns

Backend & Data (db-, auth-, api-)

Skill Description
db-postgres PostgreSQL with Drizzle ORM
db-sqlite SQLite with Prisma + Litestream backup
api-rest REST API conventions with Zod validation
auth-better-auth Better Auth integration
auth-lucia Lucia auth patterns

MCP & AI Integration

Skill Description
mcp-apps Build MCP Apps with interactive UI for Claude/ChatGPT/VSCode

Infrastructure (infra-, secrets-)

Skill Description
infra-docker Local Docker development
infra-railway Railway deployment
infra-env Environment configuration patterns
secrets-1password 1Password CLI for secrets management

Dev Environment

Skill Description
tools-mise Mise dev tool manager - tasks, tools, environments, hooks

Tools & Integrations (tools-)

Skill Description
tools-linear Linear issue tracking
tools-youtube YouTube downloads and transcripts
tools-repo-review GitHub repository analysis
tools-email Email with Resend and React Email
tools-artifacts Build HTML artifacts with React and shadcn/ui
tools-posthog Feature flags with PostHog

Document Skills

Skill Description
document-skills/docx Microsoft Word documents
document-skills/pdf PDF generation and manipulation
document-skills/pptx PowerPoint presentations
document-skills/xlsx Excel spreadsheets

What are Skills?

Skills are a simple yet powerful pattern for giving AI coding agents specialized capabilities. A skill is a Markdown file telling the model how to do something, optionally accompanied by scripts and reference documents.

Anatomy of a Skill

skills/my-skill/
├── SKILL.md           # Core instructions (required)
├── scripts/           # Helper scripts (optional)
│   └── helper.py
├── references/        # Supporting docs (optional)
│   └── api-docs.md
└── LICENSE.txt        # License (optional)
  • SKILL.md (required): Instructions for the agent. YAML frontmatter provides metadata.
  • scripts/ (optional): Pre-written scripts the agent can execute
  • references/ (optional): Supporting documentation, examples, templates

Why Skills Work

Skills leverage that modern AI agents can read files and execute commands. Any task you can accomplish by typing commands can be encoded as a skill.

The design is token-efficient: the agent only loads full skill content when needed. Skill discovery uses just the frontmatter metadata (a few dozen tokens).


Creating Your Own Skills

Automatically

Ask Cursor to create a skill:

Create a skill for managing AWS S3 buckets

The skill-creator skill guides the process.

Manually

  1. Create skills/my-skill/SKILL.md:
---
name: my-skill
description: Brief description for discovery. Explain when this skill should be used.
---

# My Skill

Detailed instructions for the agent...
  1. Add optional scripts in skills/my-skill/scripts/
  2. Add optional reference docs in skills/my-skill/references/

MCP Server Setup

For global skill access or legacy Cursor versions.

Prerequisites

  • Python 3.10+
  • uv package manager

Installation

  1. Clone this repo
  2. Copy .cursor/mcp_example.json to .cursor/mcp.json
  3. Update the path:
{
  "mcpServers": {
    "cursor-skills": {
      "command": "uv",
      "type": "stdio",
      "args": ["run", "/path/to/cursor-skills/mcp/skills_mcp.py"]
    }
  }
}
  1. Reload Cursor (Cmd+Shift+P → "Developer: Reload Window")
  2. Verify in Settings → Tools and MCP: "cursor-skills" shows 4 tools

Global Setup

To use skills across all projects:

  1. Place mcp.json in ~/.cursor/ (global config)
  2. Copy contents of .cursor/rules/main_rule.mdc to Cursor Settings → Rules

MCP Tools

Tool Description
list_skills List installed skills with descriptions
invoke_skill Load a skill's full instructions
find_skill Browse community skills directory
import_skill Import from GitHub URL

Import from GitHub (MCP)

Please import https://github.com/anthropics/skills

The import tool automatically:

  • Detects single skills vs. directories
  • Downloads all files including scripts
  • Validates SKILL.md exists
  • Skips already-installed skills

External Skill Sources

Import skills from other repositories:


Project Structure

cursor-skills/
├── .cursor/
│   ├── mcp_example.json       # MCP config template
│   └── rules/
│       └── main_rule.mdc      # Orchestrator rules (for MCP)
├── mcp/
│   └── skills_mcp.py          # MCP server
├── skills/                    # Skills collection
│   ├── ui-design-system/      # ui- prefix for frontend
│   ├── db-postgres/           # db- prefix for database
│   ├── infra-railway/         # infra- prefix for infrastructure
│   ├── tools-linear/          # tools- prefix for integrations
│   ├── feature-build/         # No prefix for core workflow
│   └── ...
├── README.md
└── LICENSE

Troubleshooting

Native Skills Not Discovered

  • Ensure skill has SKILL.md file with valid frontmatter
  • Check .cursor/skills/ directory exists
  • Restart Cursor after adding skills

MCP Server Not Starting

  • Check Cursor logs: View → Output → MCP
  • Verify Python 3.10+ installed
  • Verify uv installed: https://github.com/astral-sh/uv
  • Check paths in .cursor/mcp.json

Skills Not Found (MCP)

  • Ensure skill directory has SKILL.md file
  • Directory name must not start with . or _
  • Verify skills/ directory exists at repo root

License

This repository is licensed under the MIT License.

Third-party skills imported from external sources retain their original licenses. See individual skill directories for LICENSE files.

About

A curated set of skills for Cursor Agent use.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages