Akgithub2028

CodeContext OS

Name: CodeContext OS
Author: Akgithub2028

AI community

SOURCE https://github.com/Akgithub2028/CodeContext-OS

Description

Task-aware code intelligence for Claude Code and AI coding agents — combining AST structure, hybrid BM25 + FAISS retrieval, call-graph traversal, reranking, and token-budgeted evidence context via MCP.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

CodeContext OS

Task-Aware Code Intelligence & Advanced Context Retrieval for Claude Code & Coding Agents

[![Claude Code](https://img.shields.io/badge/Claude%20Code-Native%20MCP-CC785C?style=for-the-badge&logo=anthropic&logoColor=white)](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/overview) [![FastMCP Protocol](https://img.shields.io/badge/FastMCP-Stdio%202026-0D9488?style=for-the-badge&logo=fastapi&logoColor=white)](https://modelcontextprotocol.io) [![Python](https://img.shields.io/badge/Python-3.11%20%7C%203.12-3776AB?style=for-the-badge&logo=python&logoColor=white)](https://python.org) [![FAISS Vector Engine](https://img.shields.io/badge/Vector%20Store-FAISS%20Dense-6366F1?style=for-the-badge&logo=meta&logoColor=white)](https://github.com/facebookresearch/faiss)
[![SQLite WAL Engine](https://img.shields.io/badge/Graph%20Index-SQLite%20WAL-0284C7?style=for-the-badge&logo=sqlite&logoColor=white)](https://sqlite.org) [![LLM Judge](https://img.shields.io/badge/LLM%20Judge-NVIDIA%20120B-76B900?style=for-the-badge&logo=nvidia&logoColor=white)](https://build.nvidia.com) [![Test Suite](https://img.shields.io/badge/Tests-63%2F63%20Passed-10B981?style=for-the-badge&logo=pytest&logoColor=white)](tests/) [![License](https://img.shields.io/badge/License-MIT-F59E0B?style=for-the-badge&logo=open-source-initiative&logoColor=white)](LICENSE)

CodeContext OS Architecture & Claude Code Integration

Give Claude Code and AI coding agents a compact, task-shaped, evidence-backed view of a repository without forcing them to rediscover architecture through repeated file reads.

[Architecture](docs/architecture.md) • [RAG Deep Dive](docs/rag_deep_dive.md) • [MCP Protocol Guide](docs/mcp_guide.md) • [Empirical Benchmarks](docs/evaluation_and_benchmarks.md) • [Examples](examples/)