🔐 Code Guardian

Code Guardian is a Visual Studio Code extension that integrates AI-powered security analysis into your coding workflow. It uses Ollama to analyze JavaScript and TypeScript code for potential security vulnerabilities, bad practices, and insecure coding patterns.

---

✨ Features

🎯 Workspace-Level Security

• 🔐 Workspace Security Dashboard (NEW in v1.0.6) Comprehensive security overview with intelligent scoring (0-100), severity breakdown, and interactive visualizations.

• 📊 Security Score Calculator Automated letter grades (A-F) based on weighted vulnerability analysis normalized by codebase size.

• 📈 Vulnerability Heatmap Interactive bar chart showing issue distribution across severity levels.

🔍 File-Level Analysis

• Real-time Analysis - Automatic security checks as you type (800ms debounced)
• Full File Scanning - Complete file analysis with VS Code diagnostics integration
• AI-Powered Co-Pilot - Interactive Q&A for deep security insights

💡 Smart Features

• ⚡ Smart Caching - LRU cache providing 95-98% time reduction on repeated analysis
• 🧠 RAG Enhancement - Optional knowledge retrieval for improved accuracy
• 🔄 Dynamic Security Data (NEW in v1.0.7) - 165+ entries from NVD (100 CVEs), OWASP (complete Top 10), CWE (20 patterns), GitHub, npm
• 🛠️ Quick Fixes - One-click application of AI-suggested secure code
• 🎯 Multi-Model Support - 20+ Ollama models to choose from

🚀 Performance & Privacy

• 100% Local - All analysis happens on your machine via Ollama
• Fast - Intelligent caching and debouncing for optimal performance
• Robust - Retry logic with exponential backoff handles transient errors

---

🚀 Quick Start

Prerequisites

# 1. Install Ollama
# Visit: https://ollama.com

# 2. Pull a recommended model
ollama pull qwen2.5-coder:7b

# 3. Verify Ollama is running
ollama list

Installation

# Clone and install
git clone https://github.com/mdhafizur/code-guardian.git
cd code-guardian
npm install

# Compile and launch
npm run compile

# Press F5 in VS Code to start Extension Development Host

---

🧪 Usage

🔐 Workspace Security Dashboard (NEW)

Get a comprehensive security overview of your entire workspace:

1. Open Command Palette (Cmd/Ctrl + Shift + P)
2. Run Code Guardian: 🎯 Workspace Security Dashboard
3. View:
   • Overall security score (0-100) and grade (A-F)
   • Issues by severity (Critical, High, Medium, Low)
   • Vulnerability heatmap visualization
   • Top 20 most vulnerable files
4. Click files to navigate and fix issues
5. Rescan after fixes to see improvements

Performance: 2-5 minutes for 50-200 files on medium projects.

🔍 Real-time Analysis

Automatic security checks while you code:

• Open any JavaScript/TypeScript file
• Place cursor in a function
• Extension analyzes and highlights issues automatically
• Hover for detailed explanations

📄 Full File Analysis

Complete file scanning:

1. Open Command Palette → Code Guardian: 🔍 Analyze Full File
2. View diagnostics in Problems panel
3. Hover over issues for details and remediation

🤖 AI Security Co-Pilot

Interactive security analysis:

1. Select code or place cursor on a line
2. Run Code Guardian: 🧠 Analyze Selected Code with AI
3. View AI analysis in interactive webview
4. Ask follow-up questions
5. Switch between models as needed

Other Commands

• 📊 View Cache Statistics - Monitor cache performance
• 🎯 Select AI Model - Choose from 20+ models
• 🔄 Update Vulnerability Data - Refresh security knowledge base
• 💬 Contextual Q&A - Ask questions about your codebase

---

🧠 Supported Models

Code Guardian supports 20+ Ollama models:

Family	Models	Best For
Qwen 2.5-Coder	0.5B, 1.5B, 3B, 7B, 14B, 32B	Recommended - Best balance
CodeLlama	7B, 13B, 34B, 70B	Excellent for security analysis
DeepSeek-Coder	1.3B, 6.7B, 33B	Fast and accurate
Gemma 3	270M, 1B, 4B, 12B, 27B	Multimodal capabilities
StarCoder2	3B, 7B, 15B	Next-generation models
WizardCoder	33B	Advanced code generation
StableCode	3B	Reliable performance

Recommended Models

Use Case	Model	Size	Performance
Fast Prototyping	qwen2.5-coder:1.5b	900MB	⚡⚡⚡ Fast, ⭐⭐ Accurate
Balanced	qwen2.5-coder:7b	4.7GB	⚡⚡ Good, ⭐⭐⭐⭐ Very Accurate
Production	codellama:13b	7.3GB	⚡ Slower, ⭐⭐⭐⭐⭐ Excellent
Expert	qwen2.5-coder:32b	19GB	🐢 Slow, ⭐⭐⭐⭐⭐ Best

---

⚙️ Configuration

Configure via VS Code Settings (Cmd/Ctrl + ,):

Model Selection

{
  "codeGuardian.model": "qwen2.5-coder:7b",
  "codeGuardian.customModel": "",
  "codeGuardian.ollamaHost": "http://localhost:11434",
  "codeGuardian.autoRefreshModels": true
}

RAG Enhancement

{
  "codeGuardian.enableRAG": true
}

Enable/disable Retrieval-Augmented Generation for enhanced vulnerability detection using curated security knowledge.

---

📊 Project Structure

src/
├── extension.ts             # Extension entry point
├── analyzer.ts              # LLM analysis engine
├── analysisCache.ts         # LRU cache implementation
├── workspaceScanner.ts      # Workspace-wide scanner
├── dashboardWebview.ts      # Dashboard UI generator
├── modelManager.ts          # AI model management
├── ragManager.ts            # RAG knowledge base
├── vulnerabilityDataManager.ts  # Vulnerability data updates
├── diagnostic.ts            # VS Code diagnostics
├── actions.ts               # Quick fixes
├── functionExtractor.ts     # Code parsing
└── webview.ts               # Analysis UI

---

🏗️ Architecture & Diagrams

System Architecture

Code Guardian uses a layered architecture for optimal performance and maintainability:

6 Core Layers:

1. User Interface Layer - VS Code editor integration, status bar, quick fixes
2. Extension Core Layer - Command handling, event management, VS Code API
3. Intelligence & Analysis Layer - Security analysis, caching, RAG, workspace scanning
4. External Data Sources - NVD, OWASP, CWE, GitHub APIs (165+ vulnerability entries)
5. AI Processing Layer - Ollama LLM server with 20+ models, embeddings
6. Webview Presentation Layer - Interactive dashboards, Q&A UI, visualizations

Key Components:

• Analysis Cache (LRU): 95-98% hit rate, crypto-based hashing
• RAG Manager: Vector store with HNSWlib for semantic knowledge search
• Vulnerability Data Manager: Real-time CVE/OWASP/CWE updates (24h cache)
• Workspace Scanner: Concurrent file analysis with progress tracking
• Dashboard Generator: Security scoring (0-100, A-F), heatmap visualization

Performance Features:

• Smart caching reduces LLM calls by 95-98%
• Retry logic with exponential backoff (3 attempts)
• Debounced real-time analysis (800ms)
• Lazy loading for heavy components (RAG)

Complete Diagrams

📐 Complete Sequence Diagram

• 9 comprehensive flows covering all features
• Shows cache, RAG, vulnerability data integration
• Real-time analysis, workspace dashboard, Q&A flows
• Performance optimizations and error handling

🏛️ Complete System Architecture

• All 6 architectural layers with component details
• Data flow patterns and integration points
• Technology stack and dependencies
• Scalability considerations and future enhancements

Architecture Highlights

Analysis Flow (with all optimizations):

User Action → Cache Check (95-98% hit) →
[Cache Miss] → RAG Knowledge Search → Vulnerability Data Fetch →
Enhanced Prompt → LLM Analysis (with retry) → Cache Store →
Display Results

Workspace Dashboard Flow:

Command → File Discovery → For Each File:
  [Cache → Analyze → Store] →
Calculate Score → Generate Heatmap → Display Dashboard

---

🗺️ Roadmap

✅ v1.0.6 (Current)

• ✅ Workspace security dashboard
• ✅ Security scoring algorithm (0-100, A-F grades)
• ✅ Vulnerability heatmap visualization
• ✅ Smart caching (LRU, 95-98% faster)
• ✅ RAG enhancement
• ✅ 20+ model support
• ✅ Error handling with retry logic
• ✅ Comprehensive evaluation framework (48 test cases)

🚧 v1.1.0 (In Progress)

• 📄 Export reports (PDF, HTML, JSON, SARIF)
• 📈 Trend analysis over time
• 🧠 Custom knowledge base editor
• 🔄 CI/CD integration (GitHub Actions, GitLab CI)
• 🎯 Filter dashboard by severity

🔮 v1.2.0+ (Planned)

• 🌍 Multi-language support (Python, Java, Go, Rust, PHP)
• 👥 Team collaboration features
• 📊 Advanced visualizations (dependency graphs, attack surfaces)
• 🎯 Custom security rules engine
• 🔌 Static analysis tool integration (ESLint, Semgrep)
• ☁️ Cloud-based model support

---

📦 Packaging & Publishing

# Install VSCE
npm install -g @vscode/vsce

# Package extension
npm run package
vsce package

# Publish to marketplace (requires publisher account)
vsce publish

---

🧪 Testing & Evaluation

Run Tests

# Run all tests
npm test

# Run unit tests only
npm run test:unit

# Run with coverage
npm run test:coverage
npm run coverage:view

Evaluate Models

# Evaluate model accuracy on 48 vulnerability test cases
npm run evaluate

Benchmark Performance

# Performance benchmarking
npm run benchmark

---

📚 Documentation

For Users

• Quick Start Guide - Get up and running in minutes
• Main Documentation - Complete documentation index

For Developers

• Deployment Guide - Package and publish the extension
• Makefile Commands - Build automation (make help)
• Evaluation Framework - Model accuracy testing
• Utility Scripts - Development scripts

---

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

---

📄 License

MIT License © 2025 Md Hafizur Rahman

See LICENSE file for details.

---

🙌 Acknowledgements

• Ollama - Local LLM runtime
• VS Code Extension API - Extension framework
• TypeScript - Type-safe development
• LangChain - RAG implementation
• Marked.js - Markdown rendering

---

🐛 Support & Issues

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Repository: GitHub

---

⚠️ Disclaimer

This tool is meant for developer assistance. Always validate results manually or with expert review in production workflows. Code Guardian is not a replacement for professional security audits.

---

📈 Stats

---

🔐 Code Guardian - Your AI-Powered Security Companion

Making secure coding accessible to every developer.