Skills Library

A collection of reusable AI agent skills for research, intelligence gathering, design, and content creation. Drop a folder into your agent's skills directory and it just works.

What are skills?

Skills are structured instructions that give AI agents domain expertise and systematic workflows. Instead of prompting from scratch, you load a skill and get consistent, high-quality output. Works with OpenClaw, Claude Code, Cursor, and other LLM-based tools.

Available Skills

Research & Intelligence

Skill	Description	Auth Required
hn-search	Search & monitor Hacker News via the Algolia API	❌ Free, no key
producthunt	Search Product Hunt launches via GraphQL V2 API	✅ Free dev token (~2 min setup)
appstore-intel	App Store ratings, reviews, and metadata (iOS + Android)	❌ Free, no key
trademark-search	USPTO trademark availability search	❌ Free, no key

Agent Commerce

Skill	Description	Auth Required
privacy-cards	Create virtual cards for agent purchases via Privacy.com API	✅ Privacy.com API key

Design

Skill	Description	Auth Required
motion-design-patterns	Framer Motion patterns — springs, staggers, layout animations, micro-interactions	❌ None
design-tokens	Type scales, color palettes, spacing grids, WCAG contrast, dark mode derivation	❌ None
creative-direction	Image prompt templates, model selection, anti-generic patterns	❌ None
visual-qa	Screenshot review against design intent using vision models	❌ None

Content & Quality

Skill	Description	Auth Required
recursive-improvement	Generate → Evaluate → Improve loop for higher-quality output	❌ None

Install

Via npx (recommended)

# Install all skills
npx skills add dylanfeltus/skills

# Install a specific skill
npx skills add dylanfeltus/skills --skill visual-qa

Auto-detects your agent (Claude Code, OpenClaw, Codex, Cursor, etc.)

Manual

1. Copy the skill folder into your agent's skills directory
2. Reference the SKILL.md in your system prompt or project instructions
3. The agent follows the workflow and instructions defined in the skill

Each skill includes:

• SKILL.md — Full instructions the agent reads (trigger conditions, API details, error handling, output format)
• README.md — Human-facing docs

Contributing

More skills coming soon. Follow @dylanfeltus for updates.

---

Built by Stratus Labs