Piaegis - Next Generation LLM powered SAST & DAST

Tired of chasing down critical security vulnerabilities late in the development cycle? Our intelligent platform automatically detects SANS Top 25 and emerging threats in your production line projects.

Comment

🛡️ Piaegis: Next-Gen LLM-Powered Security Suite

DEMOS : https://drive.google.com/drive/folders/10L25hW8l-mf8dQJhoX6kWxi8Mj7X0hvY?usp=drive_link

✨ Inspiration

Security vulnerabilities are a constant headache for dev teams. We've all been there:

  • 🌙 Late nights patching critical flaws before release
  • 😰 Stress investigating potential breaches
  • 🤔 The feeling we should be doing more for security

Traditional SAST and DAST tools flood us with alerts (many false positives) and lack context-aware remediation advice.

Enter Piaegis: Leveraging AI to identify vulnerabilities more accurately and provide intelligent, actionable guidance—less like an auditor, more like a helpful security expert at your side.

🔍 What It Does

Piaegis integrates seamlessly into your development lifecycle, focusing on SANS Top 25 Most Dangerous Software Errors, OWASP Top 10, business logic flaws, and emerging threats.

Core Components:

🛡️ Piaegis Shield (SAST)

  • LLM-powered static code analysis for .NET, Java, and more
  • Understands semantic context beyond pattern matching
  • Identifies complex vulnerabilities with greater accuracy
  • Provides tailored remediation suggestions

⚔️ Piaegis Sword (DAST)

  • Integrates with OWASP ZAP for dynamic testing
  • Filters noise from raw reports
  • Presents findings in clear, actionable format
  • Provides insights on vulnerability implications

🏰 Piaegis Fortify (Interactive Security Debugger)

  • Conversational security analysis
  • Lets developers ask questions about vulnerabilities
  • Discusses security best practices
  • Offers context-aware remediation based on codebase history

🔧 How We Built It

  • Backend: Python
  • Frontend: Streamlit for user-friendly interfaces
  • LLM Integration: OpenAI, potentially Google Gemini and OpenRouter
  • DAST: OWASP ZAP API integration
  • Data Management: PostgreSQL, Neo4j, Redis + Celery
  • Containerization: Docker Compose
  • DevOps: Jenkins integration

🧗 Challenges We Faced

  • ✏️ Prompt Engineering: Balancing accuracy, minimizing false positives
  • 🧩 Diverse Codebases: Adapting to different project structures
  • 🔌 Tool Integration: Seamlessly working with ZAP and CI/CD systems
  • ⚡ Performance: Optimizing for large codebases
  • 🔄 Dependencies: Managing multiple services
  • ⚖️ Accuracy Balance: Reducing false positives without missing vulnerabilities
  • 💻 Platform-Specific Analysis: Tailoring for .NET, Java, Android, iOS

🏆 Accomplishments

  • ✅ Functional core components (Shield, Sword, Fortify)
  • 🐳 Complete Docker environment
  • 🤖 Successful LLM and DAST tool integration
  • 👨‍💻 User-friendly Streamlit interfaces
  • 🎯 Focus on actionable insights
  • 🔝 Prioritized SANS Top 25 vulnerabilities

📚 What We Learned

  • LLM strengths and limitations in security context
  • Complexity of comprehensive security testing
  • Importance of seamless workflow integration
  • Benefits of microservices architecture
  • Value of containerization

🚀 What's Next

  • 🌐 Enhanced language support (.NET, Java, Android, iOS)
  • 🎯 Improved accuracy, fewer false positives
  • 🔍 Deeper DAST analysis
  • 🧠 Business logic vulnerability detection
  • 🔭 Emerging threat detection
  • 🔧 IDE and CI/CD integrations
  • 📊 Enhanced reporting and remediation guidance
  • 🤝 Community engagement
  • 📈 Graph-based security analysis
  • ⚡ Performance optimization

Built With

Share this project:

Updates