EXBOW‑lite Autonomous Bug‑Hunting Tool

🚀 Quickstart (GitHub Codespaces)

1. Clone this repo and click Code → Open in Codespaces
2. Add your API keys by copying .env.example to .env
3. Codespace setup runs automatically

🛠️ Example flow

python recon/scan.py example.com
python ai/analyze.py
python exploit/exploit.py

Reports are generated in the reports/ folder for review. python recon/scan.py example.com python ai/analyze.py python exploit/exploit.py markdown-pdf reports/report_template.md -o reports/latest_report.pdf git add . git commit -m "Initial working version" git push origin main

EXBOW-Lite: Autonomous AI-Powered Bug Hunting System

✅ Project Summary

Goal: Build a modular AI system that automates the bug bounty kill-chain:

Recon → AI Analysis → Exploitation → Reporting

Outcome: A local or cloud-based tool that autonomously finds and reports real bugs using LLM/AI assistance.

---

🧩 Phase Roadmap

Phase	Description
0 🔐 Config & Setup	Install tools, get keys, setup environment, configure AI stack
1 📡 Recon Engine	Asset discovery + tagging using tools + AI-based classification
2 🧠 AI Vuln Analyzer	Use LLM to predict potential vulnerabilities from recon data
3 💣 Exploit Generator	Auto-generate and test PoCs using Python, curl, SQLmap, or Selenium
4 📝 Auto Reporting Engine	Generate reports with CVSS calculation; export Markdown → PDF
5 🔄 Automation & Pipelines	Loop full chain using Cron or Apache Airflow

---

⏱ Estimated Time

Phase	Time (1 Dev, Focused)
Setup & Config	1–2 days
Recon Engine	1–2 days
AI Analyzer	2–4 days
Exploit Generator	3–5 days
Reporting Engine	2–3 days
Pipeline Automation	2 days
Total	11–18 days

---

⚙️ Tech Stack

🧠 AI / LLM

Tool	Role	Notes
Groq API	LLM backend	Free, optional
LLaMA 3 / DeepSeek	Local AI runtime	For offline use
OpenAI GPT-4o	Optional premium	Backup LLM

🔍 Recon

Tool	Role
subfinder	Subdomain discovery
httpx	Live host detection
nuclei	Template-based vulnerability scanning
dirsearch	Directory brute force discovery

💣 Exploit

Tool	Role
Python Scripts	PoC generation & validation
curl / sqlmap	Automated exploitation
Selenium	Browser automation

📝 Reporting

Tool	Role
Markdown	Human-readable report
Pandoc / PDF	Export reports to PDF
GitHub CLI	Optional push reports to repo

---

🔐 Accounts & Key Config

Tool	Config Needed	How to Get
Groq API	API Key	Groq Console
OpenAI GPT	OPENAI_API_KEY	OpenAI Platform
GitHub	GITHUB_TOKEN (optional)	Developer Settings → Tokens
Nuclei	None	CLI install
Subfinder	None	CLI install

Recommended .env file structure:

OPENAI_API_KEY=sk-xxxxxxxxxxxxxxxx
GROQ_API_URL=https://api.groq.com/openai/v1/chat/completions
GITHUB_TOKEN=ghp_xxxxxxxxxxxxxxxx


---

##🚀 First Launch Plan

###1️⃣ Install Base Tools

brew install subfinder httpx nuclei dirsearch
pip install selenium requests python-dotenv

2️⃣ Install LLM Tools

Groq API:

1. Create free account: https://console.groq.com


2. Copy your API key



Ollama (Local LLM):

curl -fsSL https://ollama.com/install.sh | sh
ollama run deepseek-coder:33b

3️⃣ Clone Repository

git clone <REPO_URL>
cd <REPO_NAME>

4️⃣ Setup Environment

Create .env in the root directory with API keys

Ensure Python dependencies installed (pip install -r requirements.txt)


5️⃣ Run Initial Recon Test

python recon_bot.py -d hackerone.com

6️⃣ Review Results

live.txt → Live hosts discovered

nuclei-findings.txt → Vulnerabilities found

classified.txt → Assets classified by AI


✅ System ready for integration with AI Analyzer, Exploit Generator, and Reporting Engine.


---

🔧 Notes

Ensure your .env is never pushed to public repos

Start with small, legal targets (like your own lab or authorized bug bounty targets)

All modules are modular, so you can test each phase independently



---

📌 Future Modules

AI Vulnerability Analyzer → LLM-based prediction & prioritization

Exploit Generator → Auto-PoC & validation

Auto-Reporting → Markdown → PDF + CVSS scoring

Pipeline Automation → Cron or Airflow for full chain execution



---

🧠 Learning & Contribution

Read the docs for each tool (subfinder, nuclei, dirsearch)

Familiarize with OpenAI / Groq APIs for AI integration

Contribute to modules independently — modular design ensures easy scaling


---

I can also create a **fully functional `recon_bot.py`** next with:

- Subdomain scanning (`subfinder`)  
- Live host detection (`httpx`)  
- Nuclei vulnerability scanning  
- Smart AI classification + Groq/OpenAI prompt builder