AttackGraph AI is an AI-powered application security platform that aggregates findings from multiple security tools, correlates related vulnerabilities, generates exploitability analysis, maps attack techniques to MITRE ATT&CK, and visualizes realistic attack paths to help security teams prioritize remediation.
Modern security programs often receive thousands of findings from different scanners with little context regarding exploitability, attack chaining, or business impact.
AttackGraph AI addresses this problem by:
- Aggregating findings from multiple security scanners
- Normalizing and correlating related vulnerabilities
- Generating AI-driven exploitability assessments
- Producing remediation guidance
- Mapping attack techniques to MITRE ATT&CK
- Modeling realistic multi-stage attack paths
- Visualizing attack chains through an interactive attack graph
The platform enables security teams to move from individual findings to prioritized attack scenarios.
- Semgrep Static Application Security Testing (SAST)
- Gitleaks Secret Detection
- Unified finding ingestion pipeline
- Vulnerability normalization
- Category-based correlation
- Duplicate detection
- Attack surface aggregation
- Exploitability assessment
- Confidence scoring
- Remediation recommendations
- Attack narrative generation
- Multi-stage attack chain generation
- Risk scoring
- Likelihood assessment
- Business impact analysis
- MITRE ATT&CK mapping
- Executive security dashboard
- Attack graph visualization
- Correlated finding views
- Attack path prioritization
Developer Code
│
▼
┌─────────────┐
│ Semgrep │
└─────────────┘
│
▼
┌─────────────┐
│ Gitleaks │
└─────────────┘
│
▼
┌─────────────────┐
│ Correlation │
│ Engine │
└─────────────────┘
│
▼
┌─────────────────┐
│ Attack Path │
│ Engine │
└─────────────────┘
│
▼
┌─────────────────┐
│ OpenAI GPT-4o │
└─────────────────┘
│
▼
┌─────────────────┐
│ React Dashboard │
└─────────────────┘
- React
- TypeScript
- Axios
- ReactFlow
- FastAPI
- SQLAlchemy
- SQLite
- Semgrep
- Gitleaks
- OpenAI GPT-4o-mini
AttackGraph AI can model attack chains such as:
An attacker exploits a vulnerable SQL query, extracts credentials from the database, and escalates privileges to gain administrative access.
An exposed GitHub token provides access to source repositories containing proprietary code and additional secrets.
Compromised cloud credentials allow attackers to access and manipulate cloud resources, potentially leading to data breaches.
Cross-Site Scripting vulnerabilities enable session theft and user impersonation.
- Multi-tool vulnerability scanning
- Secret detection
- Finding correlation
- AI exploitability analysis
- AI remediation recommendations
- MITRE ATT&CK mapping
- Attack path generation
- Risk scoring
- Attack graph visualization
- Executive risk prioritization
- Trivy dependency scanning
- SBOM generation and analysis
- CVE enrichment
- Cloud attack path modeling
- Infrastructure security analysis
- Executive security summaries
- Risk trend analysis
- Automated threat modeling
- Security posture scoring
- Neo4j graph database integration
- Multi-project support
- Historical scan tracking
- Team collaboration
- Real-time monitoring
cd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reloadcd frontend
npm install
npm run devBackend API:
http://localhost:8000/docs
Frontend Dashboard:
http://localhost:5173
AttackGraph AI is intended for security research, education, and authorized security testing only.
Use only against systems and applications you own or are explicitly authorized to assess.