📖 Table of Contents

🌟 Why Use This Repository?

Pinpoint Accuracy

Skills use "Progressive Disclosure" — your AI gets only the context it needs, when it needs it. Zero bloat. Zero hallucination.

Zero-Friction Setup

Pre-packaged .skill files are ready for instant drag-and-drop deployment into any AI interface.

Self-Improving

Built-in skill-creator lets your AI autonomously write, benchmark, and optimize new skills for you.

Production-Grade

Generate stunning UI designs, beautiful generative art, and perfectly formatted corporate documents.

📦 Quick Start Installation

🟢 Option 1 — Drag & Drop (No-Code)

The fastest way to enhance your AI. No terminal. No setup. Just drag.

Step Action
1 📥 Download — Clone or download the ZIP of this repository
2 📂 Navigate — Open _claude (skill-zip) to upload directly/skill form/
3 🔍 Select — Find your desired .skill file (e.g., frontend-design.skill)
4 📤 Deploy — Drag and drop the file into claude.ai/customize/skills
5 💬 Activate — Ask your AI: "Use the frontend-design skill to create a landing page"

🔵 Option 2 — Build from Source

For developers who want to modify, inspect, or create custom skills.

# 1. Clone the repository
git clone https://github.com/Shivam990q/better-than-claude-skills.git
cd better-than-claude-skills

# 2. Navigate to the target skill directory
cd frontend-design/

# 3. Make your edits to SKILL.md, scripts/, or references/
#    Customize the skill to your exact needs

# 4. Package your custom skill using the built-in bundler
python -m scripts.package_skill path/to/skill-folder

# 5. Your new .skill file is ready for deployment! 🚀

🧰 The Skill Arsenal

📄 Document & Data Mastery

Supercharge your AI's ability to read, write, and manipulate complex file formats.

docx — Word Processing
Ultimate control over .docx files. Create, read, and manipulate Word documents with tables of contents, tracked changes, headers/footers, and flawless formatting using low-level XML manipulation and JavaScript libraries.

xlsx — Spreadsheet Processing
Advanced spreadsheet manipulation. Read, fix, format, and generate .xlsx, .xlsm, .csv, or .tsv files. Uses pandas for analysis and openpyxl for formatting. Includes mandatory formula recalculation workflows.

pdf — Document Processing
Comprehensive PDF operations using pypdf, pdfplumber, and reportlab. Extract text/tables, merge, split, watermark, encrypt, fill forms, and run OCR. Handles Unicode superscript/subscript rendering pitfalls.

pptx — Presentation Processing
Create, edit, and extract content from professional .pptx slides. Includes curated color palettes, font pairing guides, layout strategies, and automated visual QA with subagent-driven inspection.

doc-coauthoring — Collaborative Writing
A structured, 3-stage workflow for co-authoring technical docs, proposals, and specs: Context Gathering → Refinement & Structure → Reader Testing with a fresh AI to catch blind spots.

🎨 Design & Visual Arts

Elevate AI generation from generic "AI slop" to museum-quality aesthetics.

frontend-design — Production UI
Generate distinctive, production-grade frontend interfaces. Prioritizes bold aesthetics, strong typography, and precise spatial composition — never generic or "AI-looking."

algorithmic-art — Generative Art
Create stunning generative art using p5.js with seeded randomness and interactive parameters. Two-phase workflow: (1) Philosophical manifesto creation, (2) Programmatic visual expression.

canvas-design — Static Visual Art
Produce museum-quality visual art and posters as .png or .pdf. Enforces design philosophies like "Concrete Poetry," "Chromatic Language," and "Geometric Silence." Mandates expert craftsmanship.

theme-factory — Visual Theming
Apply consistent visual themes across slides, docs, and HTML pages. Choose from 10 curated themes (Ocean Depths, Midnight Galaxy, Desert Rose, etc.) or generate custom themes on the fly.

brand-guidelines — Brand Enforcement
Strictly enforce official brand colors and typography (e.g., Anthropic's #d97757 orange, Poppins/Lora fonts) on all AI-generated artifacts. Auto-fallback for missing fonts.

💻 Development & Engineering

Empower your AI to build complex applications and robust server architectures.

web-artifacts-builder — Complex Web Apps
Build multi-component React/Vite/Tailwind web artifacts with state management and routing using shadcn/ui. Includes mandatory standalone HTML bundling workflow.

webapp-testing — Automated Testing
Test local web apps using Playwright. Verify frontend functionality, debug UI, and capture screenshots. Uses "Reconnaissance-Then-Action" pattern with server lifecycle management.

claude-api — LLM API Integration
Build, debug, and optimize Claude API & Anthropic SDK integrations. Includes prompt caching, adaptive thinking strategies, model migration guides, and strict SDK vs. raw HTTP hierarchies.

mcp-builder — MCP Server Development
Authoritative guide for creating robust Model Context Protocol (MCP) servers in FastMCP (Python) or Node/TypeScript. Four-phase lifecycle: Planning → Implementation → Review/Testing → Evaluation.

🛠️ Utilities & Meta-Skills

Tools that help you communicate better and build even more skills.

skill-creator — The Crown Jewel
A meta-skill that lets the AI create new skills from scratch, modify existing ones, run evaluations, benchmark performance with variance analysis, and optimize YAML triggers for maximum accuracy.

slack-gif-creator — Animated GIFs
Generate high-quality animated GIFs optimized for Slack constraints (128×128 emoji, 480×480 message). Includes easing functions, physics simulations (bounce, shake, pulse), and PIL-based drawing.

internal-comms — Corporate Communications
Generate properly formatted corporate communications: 3P updates (Progress/Plans/Problems), company newsletters, FAQ responses, incident reports, and leadership updates.

🏗️ Progressive Disclosure Architecture

Every skill is built on a 3-layer Progressive Disclosure Architecture. The AI reads what it needs, when it needs it — nothing more.

skill-name/
├── SKILL.md                 # 📍 The core logic (< 500 lines)
│   ├── YAML frontmatter     # 🧠 Trigger definition (always loaded)
│   └── Markdown body        # ⚙️ High-level workflow instructions
└── Bundled Resources        # 📦 Loaded strictly ON-DEMAND
    ├── scripts/             # 💻 Executable Python/JS tools
    ├── references/          # 📚 Deep-dive documentation
    └── assets/              # 🖼️ Templates, icons, fonts

Layer 1

Metadata
(Always Present)

YAML frontmatter acts as a "pushy" trigger — tells the AI when to activate the skill.

Layer 2

Body Instructions
(Loaded on Activation)

SKILL.md body acts as a lightweight orchestrator — tells the AI how to work.

Layer 3

Bundled Resources
(Loaded on Demand)

scripts/ and references/ are pulled into context only when explicitly needed.

🛠️ Build Your Own Skills (Skill-Creator)

📊 Evaluation & Benchmarking

A/B Testing
Run test cases with the skill vs. without (or against older versions) simultaneously.

Quantitative Assertions
Grade outputs against verifiable hard metrics: "Did it use the correct font?", "Is the formula valid?"

Efficiency Tracking
Capture total_tokens and duration_ms for every run to monitor cost and speed.

Visual Eval Viewer
Built-in HTML viewer (generate_review.py) for side-by-side visual comparison and grading.

🤝 Contributing

# 1. Fork the repository
# 2. Create your feature branch
git checkout -b feature/AmazingSkill

# 3. Commit your changes
git commit -m 'Add some AmazingSkill'

# 4. Push to the branch
git push origin feature/AmazingSkill

# 5. Open a Pull Request 🎉

📜 License

This project is licensed under the terms specified in LICENSE.txt. Some skills may contain proprietary instructions — please review individual skill licenses where applicable.

Built for the future of agentic coding and AI-assisted workflows.

Star this repo if you find it useful!

Happy Building!