AI Agent Skills, Agents & Commands

A curated collection of 115 skills, 16 agents, and 48 commands for Claude Code and Cursor. Extend your AI coding assistant with domain expertise, autonomous workflows, and structured development processes.

What’s Inside

---

Quick Start

Install Skills with skills.sh

The fastest way to install a skill from this repo:

npx add https://github.com/wpank/ai/tree/main/skills//

Example: npx add https://github.com/wpank/ai/tree/main/skills/api/api-design

OpenClaw / Moltbot / Clawbot

npx clawhub@latest install 

Example: npx clawhub@latest install animated-financial-display

Clone the Full Repo

For access to everything — skills, agents, and commands:

git clone https://github.com/wpank/ai ~/.ai-skills

Then copy what you need into your projects (see Installation below).

---

Installation

Skills via skills.sh

Install a skill from this repo by URL:

npx add https://github.com/wpank/ai/tree/main/skills//

Example: npx add https://github.com/wpank/ai/tree/main/skills/api/api-design

OpenClaw / Moltbot / Clawbot

npx clawhub@latest install 

Example: npx clawhub@latest install animated-financial-display

Agents & Commands (Manual Copy)

Agents and commands are not published to skills.sh. Clone this repo and copy what you need into your project’s .cursor/ or .claude/ directory.

git clone https://github.com/wpank/ai ~/.ai-skills

Cursor (per-project)

From your project root:

# Skills
mkdir -p .cursor/skills
cp -r ~/.ai-skills/skills// .cursor/skills/

# Agents
mkdir -p .cursor/agents
cp -r ~/.ai-skills/agents/ .cursor/agents/

# Commands
mkdir -p .cursor/commands
cp -r ~/.ai-skills/commands/* .cursor/commands/

Cursor (global)

mkdir -p ~/.cursor/skills ~/.cursor/agents ~/.cursor/commands
cp -r ~/.ai-skills/skills// ~/.cursor/skills/
cp -r ~/.ai-skills/agents/ ~/.cursor/agents/
cp -r ~/.ai-skills/commands/* ~/.cursor/commands/

Claude Code (per-project)

From your project root:

# Skills
mkdir -p .claude/skills
cp -r ~/.ai-skills/skills// .claude/skills/

# Agents
mkdir -p .claude/agents
cp -r ~/.ai-skills/agents/ .claude/agents/

# Commands
mkdir -p .claude/commands
cp -r ~/.ai-skills/commands/* .claude/commands/

Claude Code (global)

mkdir -p ~/.claude/skills ~/.claude/agents ~/.claude/commands
cp -r ~/.ai-skills/skills// ~/.claude/skills/
cp -r ~/.ai-skills/agents/ ~/.claude/agents/
cp -r ~/.ai-skills/commands/* ~/.claude/commands/

---

Skills

115 modular AI agent skills organized across 17 categories. Each skill is a self-contained SKILL.md with patterns, code examples, anti-patterns, and decision frameworks.

API (8 skills)

Designing, securing, and scaling APIs.

Backend (12 skills)

Server-side architecture, databases, and infrastructure patterns.

Design Systems (10 skills)

Token architecture, theming, component patterns, and visual design.

DevOps (6 skills)

Containers, orchestration, monitoring, and build systems.

Extraction (1 skill)

Frontend (15 skills)

React, Next.js, mobile, styling, and component architecture.

Marketing (8 skills)

Copy, conversion optimization, content strategy, and metrics.

Meta (17 skills)

Orchestration skills that coordinate other skills into complete workflows.

Product (1 skill)

Realtime (4 skills)

WebSockets, SSE, event streaming, and resilient connections.

Refinement (1 skill)

Testing (9 skills)

Quality assurance, code review, debugging, and standards.

Tools (9 skills)

CLI tools, skill management, session lifecycle, and releases.

Writing (9 skills)

Technical writing, diagrams, prompts, and communication.

---

Agents

16 autonomous workflow agents that orchestrate skills and commands into end-to-end processes. Each agent is a multi-phase workflow defined in AGENT.md.

---

Commands

48 slash commands organized in 8 groups. Each command is a markdown file with usage, examples, and implementation steps.

Bootstrap (3)

Development (31)

Docs (3)

Documentation (1)

Extraction (1)

Marketing (2)

Refinement (2)

Skills (5)

Common Workflows

# New project
/bootstrap-project → /bootstrap-design-system → /bootstrap-docs

# Feature end-to-end
/new-feature → /create-component → /create-api-route → /test-feature → /review-code

# UI development
/design-ui → /create-component → /accessibility-audit → /check-performance

# Backend service
/create-service → /create-migration → /create-api-route → /test-feature

# Ship it
/check-performance → /security-review → /deploy staging → /deploy production

# Session lifecycle
/intent → /start-task → /workflow → /complete-task → /session-summary

---

Repo Structure

.
├── skills/                  # 115 skills organized by category
│   ├── ai-chat/             #   AI chat interfaces and streaming
│   ├── api/                 #   API design, auth, caching
│   ├── backend/             #   Architecture, databases, Node.js, Go
│   ├── design-systems/      #   Tokens, theming, components
│   ├── devops/              #   Docker, K8s, Prometheus, Turborepo
│   ├── documentation/       #   ADRs, PRDs, docs organization
│   ├── extraction/          #   Pattern mining
│   ├── frontend/            #   React, Next.js, Expo, Tailwind
│   ├── infrastructure/     #   Docker env, K8s, Prometheus observability
│   ├── marketing/           #   Copy, CRO, content strategy
│   ├── meta/                #   Orchestration and workflow skills
│   ├── product/             #   Startup metrics
│   ├── realtime/            #   WebSocket, SSE, Kafka, Redis
│   ├── refinement/          #   Skill consolidation
│   ├── testing/             #   Testing, review, debugging
│   ├── tools/               #   CLI, releases, session management
│   └── writing/             #   Technical writing, diagrams, prompts
├── agents/                  # 16 autonomous workflow agents
│   ├── api/
│   ├── backend/
│   ├── bootstrap/
│   ├── content/
│   ├── debugging/
│   ├── deployment/
│   ├── design-system/
│   ├── development/
│   ├── extraction/
│   ├── frontend/
│   ├── marketing/
│   ├── migration/
│   ├── performance/
│   ├── refactoring/
│   ├── refinement/
│   └── testing/
├── commands/                # 48 slash commands
│   ├── bootstrap/           #   Project scaffolding (3)
│   ├── development/         #   Day-to-day workflows (31)
│   ├── docs/                #   Documentation artifacts (3)
│   ├── documentation/       #   Doc generation (1)
│   ├── extraction/          #   Pattern mining (1)
│   ├── marketing/           #   CRO and copy (2)
│   ├── refinement/          #   Skill promotion (2)
│   └── skills/              #   Skill management (5)
└── .claude/                 # Active installation for Claude Code
    ├── skills/              #   Runtime-loaded skills
    ├── agents/              #   Runtime-loaded agents
    └── commands/            #   Runtime-loaded commands

---

Skill Anatomy

Each skill is a directory with a required SKILL.md and optional supporting files:

skill-name/
├── SKILL.md              # Required — patterns, examples, anti-patterns
├── README.md             # Human-readable overview
├── references/           # Deep-dive docs and supporting material
├── templates/            # Starter templates and scaffolds
├── data/                 # Data files (CSV, JSON)
└── scripts/              # Automation scripts

The SKILL.md contains:

• YAML frontmatter — name, model tier, description, trigger keywords
• When to Use — scenarios where the skill applies
• Patterns — code examples, decision frameworks, checklists
• Anti-patterns — common mistakes to avoid
• Related Skills — links to complementary skills

---

Agent Anatomy

Each agent is defined in AGENT.md with:

1. Skills — Domain knowledge loaded from skills/
2. Commands — Slash commands from commands/ for sub-workflows
3. Workflow — Multi-phase process with checkpoints and quality gates
4. Model Tiers — Per-phase model recommendations (fast/standard/reasoning)

---

Contributing

Creating a New Skill

1. Create a directory under the appropriate category in skills/
2. Write a SKILL.md following the structure above
3. Add a README.md with a human-readable overview
4. Use the skill-creator skill or create-skill command for guidance
5. Validate with the skill-judge skill (120-point scoring)

Creating a New Agent

1. Create a directory under agents/
2. Write an AGENT.md with YAML frontmatter, skills list, and phased workflow
3. Add a README.md
4. Reference existing skills and commands in the workflow

Creating a New Command

1. Create a .md file under the appropriate group in commands/
2. Include: Usage, Examples, When to Use, Implementation Steps, Output format

---

License

MIT