Frun1753/claude-codecraft-toolkit

A Production-Ready Framework for Intelligent Code Quality Assurance, Vulnerability Detection, and Dependency Governance
## Overview
This repository introduces a context-aware, multi-model AI toolkit designed to transform how development teams handle code reviews, security audits, and dependency management. Unlike traditional static analysis tools that rely on predefined rules, this system leverages Claude API and OpenAI API integration to understand code semantics, project context, and architectural patterns. The result is a responsive, adaptive assistant that evolves with your codebase, reducing review cycles from days to minutes while catching issues that conventional tools miss.
The toolkit operates as a skill-based orchestration layer that can be invoked directly from your terminal, integrated into CI/CD pipelines, or deployed as a standalone service. It supports multilingual code analysis across 30+ programming languages and provides 24/7 continuous monitoring capabilities for teams operating across time zones.
## Key Features
Context-Aware Code Reviews — Understands your project's architecture, coding standards, and domain-specific requirements rather than applying generic rules
Zero-Day Vulnerability Detection — Uses AI pattern recognition to identify security flaws that signature-based scanners cannot detect
Intelligent Dependency Governance — Analyzes dependency trees for license compliance, known CVEs, and compatibility risks
Multi-Model AI Backend — Seamlessly switches between Claude API and OpenAI API based on task complexity and cost optimization
Responsive CLI Interface — Designed for both human interaction and automated pipeline execution
Continuous Learning — Adapts to your team's feedback and coding patterns over time
Enterprise-Grade Security — All code analysis occurs locally; only anonymized patterns are shared with AI APIs
## How It Works: System Architecture
mermaid
graph TD
A[Developer Push/PR] --> B[Git Hook / CI Trigger]
B --> C{Skill Selector}
C --> D[Code Review Skill]
C --> E[Security Audit Skill]
C --> F[Dependency Check Skill]
D --> G[Claude API / OpenAI API]
E --> G
F --> G
G --> H[Pattern Analysis Engine]
H --> I[Context Repository]
I --> J[Anomaly Detection]
J --> K[Report Generator]
K --> L[Markdown Report]
K --> M[Inline PR Comments]
K --> N[Slack/Email Alerts]
D --> O[Learning Feedback Loop]
O --> P[Project-Specific Model]
P --> D
The architecture follows a skill-based routing pattern where incoming code changes are classified and dispatched to specialized analysis modules. Each skill maintains its own context cache and learning model, ensuring that feedback from one review improves future analyses without cross-contamination between different skill domains.
## Supported Systems
| Operating System | Status | Notes |
| Linux (Ubuntu 20.04+) | ✅ Full Support | Recommended for production |
| macOS (Ventura+) | ✅ Full Support | Native ARM64 support |
| Windows 10/11 | ✅ Full Support | WSL2 recommended |
| FreeBSD | 🟡 Beta | Limited dependency analysis |
| Alpine Linux | 🟡 Beta | Minimal footprint deployment |
## Installation
### Prerequisites
Python 3.10+ or Node.js 18+
API keys for Claude and/or OpenAI