💡 Summary
A collection of AI skills for coding assistants, enabling various functionalities across multiple platforms.
🎯 Target Audience
🤖 AI Roast: “Powerful, but the setup might scare off the impatient.”
Risk: Medium. Review: shell/CLI command execution; outbound network access (SSRF, data egress); API keys/tokens handling and storage; filesystem read/write scope and path traversal; dependency pinning and supply-chain risk. Run with least privilege and audit before enabling in production.
AI Agent Skills
A collection of portable skills for AI coding assistants. Works with all major AI clients that support the Agent Skills Standard.
Supported AI Clients
| Client | Skills Directory | Documentation |
|--------|-----------------|---------------|
| Claude Code | ~/.claude/skills/ or .claude/skills/ | docs |
| Gemini CLI | ~/.gemini/skills/ or .gemini/skills/ | docs |
| Google Antigravity | ~/.gemini/antigravity/skills/ or .agent/skills/ | docs |
| Cursor | ~/.cursor/skills/ or .cursor/skills/ | docs |
| OpenAI Codex CLI | ~/.codex/skills/ or .codex/skills/ | docs |
| Goose | ~/.config/goose/skills/ or .goose/skills/ | docs |
Available Skills
| Skill | Description | |-------|-------------| | postgres | Read-only PostgreSQL queries with defense-in-depth security | | imagen | AI image generation using Google Gemini (cross-platform) | | deep-research | Autonomous multi-step research using Gemini Deep Research Agent | | outline | Search, read, and manage Outline wiki documents | | jules | Delegate coding tasks to Google Jules AI agent (async bug fixes, docs, tests, features) |
Google Workspace Skills
Lightweight alternatives to the full Google Workspace MCP server. Each skill has standalone OAuth authentication with cross-platform token storage via keyring.
⚠️ Requires Google Workspace account. Personal Gmail accounts are not supported. These skills use the same OAuth infrastructure as the official Google Workspace MCP.
| Skill | Description | |-------|-------------| | google-chat | List spaces, send messages, DMs, create spaces | | google-docs | Create, read, edit Google Docs | | google-sheets | Read spreadsheets, get ranges, find sheets | | google-slides | Read presentations, get text and metadata | | google-drive | Search files, list folders, download files | | google-calendar | Events, scheduling, free time lookup | | gmail | Search, read, send emails, manage labels |
Installation
Option 1: Clone entire repository
# Clone to your preferred skills directory git clone https://github.com/sanjay3290/ai-skills.git ~/.claude/skills/ai-skills # Or for other clients: # git clone https://github.com/sanjay3290/ai-skills.git ~/.gemini/skills/ai-skills # git clone https://github.com/sanjay3290/ai-skills.git ~/.gemini/antigravity/skills/ai-skills # git clone https://github.com/sanjay3290/ai-skills.git ~/.cursor/skills/ai-skills # git clone https://github.com/sanjay3290/ai-skills.git ~/.codex/skills/ai-skills # git clone https://github.com/sanjay3290/ai-skills.git ~/.config/goose/skills/ai-skills
Option 2: Copy individual skills
# Example: Install just the postgres skill cp -r skills/postgres ~/.claude/skills/
Option 3: Symlink for development
# Symlink skills for easy updates ln -s /path/to/ai-skills/skills/postgres ~/.claude/skills/postgres ln -s /path/to/ai-skills/skills/imagen ~/.claude/skills/imagen
Client-Specific Installation
Claude Code
# Global installation cp -r skills/* ~/.claude/skills/ # Or project-level cp -r skills/* .claude/skills/
Gemini CLI
# Global installation cp -r skills/* ~/.gemini/skills/ # Or workspace-level cp -r skills/* .gemini/skills/
Google Antigravity
# Global installation cp -r skills/* ~/.gemini/antigravity/skills/ # Or workspace-level cp -r skills/* .agent/skills/
Cursor
# Global installation cp -r skills/* ~/.cursor/skills/ # Or project-level cp -r skills/* .cursor/skills/
OpenAI Codex CLI
# Global installation cp -r skills/* ~/.codex/skills/ # Or repository-level cp -r skills/* .codex/skills/
Goose
# Global installation cp -r skills/* ~/.config/goose/skills/ # Or project-level cp -r skills/* .goose/skills/
Skill Setup
Each skill may require additional configuration:
Postgres
Create connections.json in the skill directory with your database credentials. See postgres/README.md.
Imagen & Deep Research
export GEMINI_API_KEY=your-api-key
Get a free key at Google AI Studio.
Note: Deep Research tasks take 2-10 minutes and cost $2-5 per query.
Outline
export OUTLINE_API_KEY=your-api-key export OUTLINE_API_URL=https://your-wiki.example.com/api # Optional
Get your API key from your Outline wiki settings.
Jules
# Install CLI (one-time) npm install -g @google/jules # Authenticate (opens browser) jules login
Connect your GitHub repos at jules.google.com. Jules works asynchronously - create a task, it runs in the background, then pull results when complete.
Google Workspace Skills
Each Google Workspace skill requires the keyring library and first-time authentication:
# Install dependency (one-time) pip install keyring # Authenticate for the skill you need (opens browser) python ~/.claude/skills/ai-skills/skills/google-chat/scripts/auth.py login python ~/.claude/skills/ai-skills/skills/google-docs/scripts/auth.py login python ~/.claude/skills/ai-skills/skills/google-sheets/scripts/auth.py login python ~/.claude/skills/ai-skills/skills/google-slides/scripts/auth.py login python ~/.claude/skills/ai-skills/skills/google-drive/scripts/auth.py login python ~/.claude/skills/ai-skills/skills/google-calendar/scripts/auth.py login python ~/.claude/skills/ai-skills/skills/gmail/scripts/auth.py login
Tokens stored securely via system keyring:
- macOS: Keychain
- Windows: Windows Credential Locker
- Linux: Secret Service API (GNOME Keyring, KDE Wallet, etc.)
Usage
Once installed, skills activate automatically based on your requests. Just ask naturally:
Postgres
- "Query my production database for active users"
- "Show me the schema of the orders table"
- "How many signups last week?"
Imagen
- "Generate an image of a sunset over mountains"
- "Create an app icon for my weather app"
- "I need a hero image for my landing page"
Deep Research
- "Research the competitive landscape of EV batteries"
- "Compare React, Vue, and Angular frameworks"
- "What are the latest developments in Kubernetes?"
Outline
- "Search the wiki for deployment guide"
- "Read the onboarding documentation"
- "Create a new wiki page for the API spec"
Jules
- "Have Jules fix the authentication bug in src/auth.js"
- "Delegate adding unit tests to Jules"
- "Ask Jules to add documentation to the API module"
- "Check my Jules sessions" / "Pull the results from Jules"
Google Workspace
- "List my Google Chat spaces" / "Send a message to Project Alpha"
- "Create a new Google Doc about the project proposal"
- "Get the content of my Q4 Budget spreadsheet"
- "What's on my calendar tomorrow?"
- "Search my Gmail for invoices from last month"
- "Find files named 'report' in my Drive"
Skill Structure
All skills follow the Agent Skills Standard:
skill-name/
├── SKILL.md # Required: Instructions for the AI agent
├── README.md # Human documentation
├── requirements.txt # Dependencies (if any)
├── .env.example # Environment variable template
└── scripts/ # Executable scripts
└── main.py
The SKILL.md file uses YAML frontmatter:
--- name: skill-name description: "When to use this skill" --- # Instructions for the AI agent
Contributing
- Fork this repository
- Create a new skill in
skills/your-skill-name/ - Include
SKILL.mdwith proper frontmatter - Add documentation in
README.md - Submit a pull request
Credits
- Google Workspace Skills - Based on the official Google Workspace MCP server by the Gemini CLI team. Uses their OAuth cloud function for authentication.
License
Apache-2.0
Pros
- Supports multiple AI clients
- Offers a variety of useful skills
- Easy installation and setup
- Cross-platform compatibility
Cons
- Requires specific client configurations
- Some skills need additional API keys
- Documentation could be more detailed
- Limited to read-only for PostgreSQL
Related Skills
banana-skills
B“Powerful, but the setup might scare off the impatient.”
pytorch
S“It's the Swiss Army knife of deep learning, but good luck figuring out which of the 47 installation methods is the one that won't break your system.”
agno
S“It promises to be the Kubernetes for agents, but let's see if developers have the patience to learn yet another orchestration layer.”
Disclaimer: This content is sourced from GitHub open source projects for display and rating purposes only.
Copyright belongs to the original author sanjay3290.
