skills.rest
VoThinhVuong
VoThinhVuongCOMMUNITY
Explore all skills by this author →
GitHub

context-extractor

Extract business logic and architectural context from source code repositories into vector embeddings.

Updated May 18, 2026
Search Tags:#code analysis#vector embeddings#code repository#business logic extraction#AI coding agents
One-click install
npx skills add https://github.com/VoThinhVuong/context-extractor-ts --skill context-extractor-vothinhvuong
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: context-extractor
Source: https://github.com/VoThinhVuong/context-extractor-ts/tree/main
Command: npx skills add https://github.com/VoThinhVuong/context-extractor-ts --skill context-extractor-vothinhvuong

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @langchain/anthropic, @langchain/core, @langchain/google-genai, @langchain/langgraph, dotenv, pg, tiktoken, tree-sitter, tree-sitter-go, tree-sitter-java, tree-sitter-javascript, tree-sitter-python, tree-sitter-typescript, zod, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill automates the process of analyzing code repositories, extracting business logic, and storing them as vector embeddings, which can then be used by AI coding agents to understand codebases efficiently.

Core Features & Use Cases

  • Code Analysis: Extracts function signatures, class definitions, and top-level declarations from source code files.
  • LLM Analysis: Sends code skeletons to an LLM to extract business rules and architectural boundaries.
  • Summarization: Generates a markdown document summarizing the extracted insights and stores them as vector embeddings.
  • Use Case: With this Skill, you can quickly generate a comprehensive overview of a codebase, which is especially useful for new team members or for understanding the business logic within a large code repository.

Quick Start

Run the context-extractor skill on your code repository by executing 'npm start' within the repository directory.

Frequently Asked Questions about context-extractor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract business logic from a code repository for AI coding agents?▼

You can extract business logic from a code repository by running this Skill, which uses tree-sitter to parse function signatures and class definitions, then sends code skeletons to an LLM to distill business rules into vector embeddings for semantic retrieval.

What programming languages does the code analysis support for generating vector embeddings?▼

The code analysis supports TypeScript, JavaScript, Python, Go, and Java source code files, parsing them via tree-sitter to extract top-level declarations and architectural boundaries for embedding generation.

Do I need a PostgreSQL database with pgvector to extract architectural context from source code?▼

Yes, a PostgreSQL database with the pgvector extension is required to store the generated markdown summaries as vector embeddings after extracting architectural context from your source code.

How does LLM analysis work when extracting business rules from a codebase?▼

LLM analysis for extracting business rules works by sending parsed code skeletons to an LLM, which identifies business rules and architectural boundaries, then generates a markdown document summarizing the insights stored as vector embeddings.

Can I use LangChain and tree-sitter to summarize a codebase for new team members?▼

Yes, this Skill integrates LangChain and tree-sitter to quickly generate a comprehensive markdown overview of a codebase, making it useful for onboarding new team members and understanding large repositories.

Related Skills

Explore More Software Engineering Skills→
xberg-ioxberg-io
26

managing-parsers

Manage tree-sitter parser cache with prefetch, list, inspect, and clean commands.

Official
Intermediate
discountedcookiediscountedcookie

postgres-vectors

Store 384-dimensional embeddings in Postgres and query cosine similarity with pgvector.

Community
Advanced
PriyanshKuniyalPriyanshKuniyal

langchain-architecture

Integrate LangChain 1.x and LangGraph for stateful LLM agent orchestration.

Community
Advanced
gatovillanogatovillano
5

code_tools

Search codebases semantically and analyze Python files for complexity and maintainability.

Community
Intermediate