Autonomous Security Operations Multi-LLM Engine

Autonomous Pentesting Powered by Agentic Cyber AI

HackGPT is an enterprise penetration testing automation suite. It orchestrates multi-phase reconnaissance, automated vulnerability scanning, safe exploit verification, ML zero-day anomaly detection, and executive compliance reporting.

Try AI Threat Simulator Installation & Usage View Source

6-Phase Workflow Recon to Executive PDF

Multi-Provider AI OpenAI, Llama, Ollama, LiteLLM

100% Safe Execution No Shell Injection Risk

guest@hackgpt:~$ python3 advance_hackgpt.py --target production-api.net --api --web
[+] HackGPT Enterprise Framework v2026.07 initialized
[+] Database connection active (PostgreSQL cluster)
[i] Multi-LLM Router: Provider initialized [OpenAI GPT-4o / Ollama Fallback]
[+] Phase 1: Reconnaissance & Subdomain Enumeration starting...
    - Discovering subdomains: api.production-api.net, admin.production-api.net
    - Port scan complete: 80/tcp (HTTP), 443/tcp (HTTPS), 8443/tcp (K8s API) open
[+] Phase 2: Vulnerability Enumeration...
    - Found unauthenticated endpoint: /api/v1/auth/tokens
    - [CRITICAL] SQL Injection flaw detected on endpoint parameter 'user_id'
[+] Phase 3: AI Threat Analysis & Payload Generation running...
    - AI Agent generated non-destructive PoC verification query
[+] Phase 4: Exploitation Verification successful. Administrator hash exposed!
[+] Phase 5: DBSCAN Log Clustering & Anomaly Detection active...
    - Flagged 3 abnormal admin API requests (Confidence: 98.4%)
[+] Phase 6: Executive PDF and HTML Audit Reports generated!
[✓] Scan finished in 14.8s. Session UUID: e41b8a92-7f21-49b0-9f12-00ab198fa391
_

1 Critical CVE

3 High Risk

98% Compliance Score

14.8s Scan Time

Target Risk Assessment: HIGH SEVERITY

OWASP A03:2021 SQL Injection Verified

Core Architecture

The 6-Phase Automated Pentesting Pipeline

HackGPT systematically executes security testing workflows like an experienced ethical hacker, powered by AI threat modeling.

PHASE 01 // INTELLIGENCE GATHERING

Passive & Active Reconnaissance

Discovers attack surfaces, identifies live subdomains, performs DNS enumeration, and catalogs exposed web ports and services using integrated tools.

Python Module: ai_engine/providers.py & cloud/service_registry.py

Key Capabilities: Subdomain discovery, Banner Grabbing, Port Scanning, Service Mapping

CLI Command Simulation

python3 advance_hackgpt.py --target company.com --recon --verbose

Live Demonstration

HackGPT AI Threat Threat Analyzer

Experience how HackGPT analyzes vulnerability patterns and generates structured remediation guidelines in real time.

Input Code Snippet / Finding Description:

AI Analysis Report Output: READY

Click "Analyze Vulnerability with HackGPT AI" or select a preset to generate real-time AI security breakdown.

Feature Suite

Built For Modern Security Engineers

Engineered with modular Python design patterns, enterprise database support, and strict ethical safety controls.

Multi-Provider AI Orchestration

Supports OpenAI GPT-4o, Anthropic Claude, offline Llama-3 (Ollama/LM Studio), and LiteLLM unified model router with automatic fallback logic.

ai_engine/providers.py

Machine Learning 0-Day Detector

Leverages DBSCAN density clustering and Isolation Forest anomaly algorithms to identify unknown zero-day behavior patterns from logs.

exploitation/zero_day_detector.py

Safe Exploitation Engine

Validates findings safely without using risky shell=True subprocess calls, avoiding unintended side-effects on target hosts.

exploitation/advanced_engine.py

Cloud & Infrastructure Security

Automates security auditing for Docker containers, Kubernetes API RBAC, load balancers, and microservice service registries.

cloud/docker_manager.py

Enterprise Dashboard & Web UI

Monitor real-time pentest progress, live vulnerability findings, interactive web consoles, and SOC logs from a modern Web App.

reporting/realtime_dashboard.py

Automated Compliance PDF Reports

Generates executive PDF and HTML audit reports mapped directly to OWASP Top 10, NIST SP 800-53, ISO 27001, and PCI-DSS v4.0.

reporting/dynamic_reports.py

Getting Started

Get Started in 3 Simple Steps

HackGPT runs on Linux/macOS/Windows with Python 3.8+ or inside Docker containers.

Clone & Install Dependencies

git clone https://github.com/yashab-cyber/HackGpt.git
cd HackGpt
pip install -r requirements-dev.txt

Configure Environment (Optional)

cp .env.example .env
# Add OPENAI_API_KEY or configure local Ollama endpoint

Launch HackGPT Assessment

# Run full pentest with Web Dashboard
python advance_hackgpt.py --target target-domain.com --api --web

Interactive Sandbox

HackGPT Interactive Shell Simulator

Test HackGPT CLI commands directly in your browser. Type help to list available commands.

hackgpt-shell-v2026.07 (bash)

Welcome to HackGPT Shell Simulator v2026.07.beta.4

Type help to view available security commands. Try: run, ai-analyze, zero-day, or compliance.

guest@hackgpt:~$

Lead Developer

Meet the Developer

Yashab Alam is the Lead Developer & Creator of the HackGPT open-source platform. With a passion for AI security, ethical hacking, and enterprise vulnerability management, Yashab actively maintains the HackGPT suite.

Primary Email yashabalam9@gmail.com

Secondary Email yashabalam707@gmail.com