CodeTree Skill

Query any code repository using natural language. Built on vectorless RAG - uses LLM reasoning instead of vector similarity.

Prerequisites

1. CodeTree must be installed:

   cd /path/to/Oh-Code-Rag
   pip install -e .

1. Set your LLM API key:

   export OPENAI_API_KEY="sk-..."
   # or ANTHROPIC_API_KEY

Commands

Index a Repository

Before querying, you need to index the repository:

./scripts/codetree.sh index /path/to/repo

This creates a .codetree/index.json in the repo.

Query Code

Ask questions about the codebase:

./scripts/codetree.sh query /path/to/repo "How does authentication work?"

Show Code Tree

Display the structure of the codebase:

./scripts/codetree.sh tree /path/to/repo

Find Symbol

Find all references to a function, class, or variable:

./scripts/codetree.sh find /path/to/repo "UserService"

Get Stats

Show repository statistics:

./scripts/codetree.sh stats /path/to/repo

Usage Examples

Example 1: Understand a New Project

# First, index the project
./scripts/codetree.sh index ~/projects/my-app

# Ask about architecture
./scripts/codetree.sh query ~/projects/my-app "What's the overall architecture?"

# Ask about specific functionality
./scripts/codetree.sh query ~/projects/my-app "How does the payment system work?"

Example 2: Code Investigation

# Find all usages of a function
./scripts/codetree.sh find ~/projects/my-app "processOrder"

# Trace the code flow
./scripts/codetree.sh query ~/projects/my-app "What happens when a user clicks checkout?"

Example 3: Onboarding Help

# Show the project structure
./scripts/codetree.sh tree ~/projects/my-app

# Get statistics
./scripts/codetree.sh stats ~/projects/my-app

# Ask where to add new features
./scripts/codetree.sh query ~/projects/my-app "Where should I add a new API endpoint?"

Supported Languages

• Python (.py)
• JavaScript (.js, .jsx)
• TypeScript (.ts, .tsx)
• Go (.go)
• Rust (.rs)
• Java (.java)

Notes

• The index is saved in .codetree/index.json inside each repo
• Re-run index after significant code changes
• Query time depends on LLM provider latency (typically 2-10 seconds)
• For large repos, first indexing may take a few minutes