Weco AI Skill

AI-powered code optimization skill for Claude Code and Cursor.

What is this?

This skill teaches your AI coding assistant how to use Weco for code optimization. When you ask to "make this faster" or "improve accuracy", the assistant will:

1. Analyze your code and environment
2. Set up an evaluation benchmark
3. Run Weco's optimization engine
4. Present results with a detailed report
5. Apply changes (with your approval)

Installation

Recommended: Weco CLI

The easiest way to install is via the Weco CLI, which sets up the skill and trigger rules for your agent:

pip install weco
weco install cursor   # For Cursor
weco install claude   # For Claude Code

Cursor Plugin

Install directly from the Cursor marketplace:

1. Open Cursor
2. Go to Settings > Plugins
3. Search for "weco"
4. Click Install

Quick install: npx

Install the skill using the Agent Skills CLI:

npx skills add weco-ai/weco-skill

This installs the skill directory (SKILL.md, CLAUDE.md, references, assets). For Cursor, this relies on the skill description for activation — for the always-on trigger rule, use weco install cursor instead.

Manual install: install.sh

./install.sh

The installer will ask which AI assistant you use (Claude Code, Cursor, or both) and set up the skill with trigger rules accordingly.

What Gets Installed

Claude Code:

~/.claude/skills/weco/
├── CLAUDE.md          # Trigger snippet (Claude reads this)
├── SKILL.md           # Full optimization workflow
├── references/        # Advanced documentation
│   ├── benchmarking.md
│   ├── ml-evaluation.md
│   └── ...
└── assets/            # Template evaluation scripts
    ├── evaluate-speed.py
    ├── evaluate-accuracy.py
    └── ...

Cursor (via weco install cursor or install.sh):

~/.cursor/
├── rules/
│   └── weco.mdc           # Always-on trigger rule
└── skills/
    └── weco/
        ├── SKILL.md       # Full optimization workflow
        ├── references/    # Advanced documentation
        └── assets/        # Template evaluation scripts

Usage

Once installed, just ask your AI assistant to optimize code:

• "Make this function faster"
• "Optimize this for speed"
• "Improve the accuracy of this model"
• "Reduce the latency"

The skill offers two modes:

• Vibe Mode: Minimal questions, maximum action. Great for simple optimizations.
• Assistant Scientist Mode: Collaborative, educational. Best for complex cases or learning.

Requirements

• Weco CLI installed and authenticated
• Claude Code or Cursor

Files

weco-skill/
├── .cursor-plugin/       # Cursor plugin manifest
│   └── plugin.json
├── SKILL.md              # Full skill instructions (source of truth)
├── CLAUDE.md             # Trigger snippet for Claude Code (ships with skill)
├── install.sh            # Interactive installer
├── README.md             # This file
├── rules/                # Cursor plugin rules
│   └── weco.mdc          # Always-on trigger rule for Cursor
├── snippets/             # Trigger snippets (used by weco-cli installer)
│   ├── claude.md         # Claude Code trigger
│   ├── claude-global.md  # Claude Code global trigger
│   └── cursor.md         # Cursor trigger (.mdc rule)
├── references/           # Advanced documentation
│   ├── benchmarking.md
│   ├── ml-evaluation.md
│   ├── gpu-profiling.md
│   └── ...
└── assets/               # Template evaluation scripts
    ├── evaluate-speed.py
    ├── evaluate-accuracy.py
    └── ...

Learn More

• Weco Documentation
• Claude Code
• Cursor