AI Summary: Claude Bug Bounty is an AI-powered agent harness tailored for professional bug bounty hunting, enabling users to streamline their testing processes. It intelligently orchestrates over 25 tools, remembers past findings across targets, validates vulnerabilities to reduce false positives, and generates ready-to-submit reports in under a minute. Notable features include Burp MCP integration for live traffic monitoring and an autonomous hunting mode that simplifies the testing workflow.

README

Claude Bug Bounty

The AI-Powered Agent Harness for Professional Bug Bounty Hunting

Your AI copilot that sees live traffic, remembers past hunts, and hunts autonomously.

Quick Start  |  How It Works  |  Commands  |  What’s New  |  Install

  13 commands  ·  7 AI agents  ·  8 skill domains
  20 web2 vuln classes  ·  10 web3 bug classes
  Burp MCP  ·  HackerOne MCP  ·  Autonomous Mode

The Problem

Most bug bounty toolkits give you a bag of scripts. You still have to:

• Figure out what to test and in what order
• Waste hours on false positives that get rejected
• Write reports from scratch every time
• Forget what worked on previous targets
• Context-switch between 15 different terminal windows

The Solution

Claude Bug Bounty is an agent harness — not just scripts. It reasons about what to test, validates findings before you waste time writing them up, remembers what worked across targets, and generates reports that actually get paid.

Quick Start

Step 1 — Install

git clone https://github.com/shuvonsec/claude-bug-bounty.git
cd claude-bug-bounty
chmod +x install.sh && ./install.sh

Step 2 — Hunt

claude                          # Start Claude Code

/recon target.com               # Discover attack surface
/hunt target.com                # Test for vulnerabilities
/validate                       # Check finding before writing
/report                         # Generate submission-ready report

Step 3 — Go Autonomous (new in v3)

/autopilot target.com --normal  # Full autonomous hunt loop
/intel target.com               # Fetch CVE + disclosure intel
/resume target.com              # Pick up where you left off

Or run tools directly — no Claude needed:

python3 tools/hunt.py --target target.com
./tools/recon_engine.sh target.com
python3 tools/intel_engine.py --target target.com --tech nextjs

How It Works

                         YOU
                          |
                    ┌─────▼─────┐
                    │   Claude   │ ◄── Burp MCP (sees your traffic)
                    │   Code     │ ◄── HackerOne MCP (program intel)
                    └─────┬─────┘
                          |
          ┌───────────────┼───────────────┐
          |               |               |
    ┌─────▼─────┐  ┌──────▼──────┐  ┌────▼────┐
    │   Recon    │  │    Hunt     │  │ Report  │
    │   Agent    │  │   Engine    │  │ Writer  │
    └─────┬─────┘  └──────┬──────┘  └────┬────┘
          |               |               |
    subfinder        scope check      H1/Bugcrowd
    httpx            vuln test        Intigriti
    katana           validate         Immunefi
    nuclei           chain A→B→C      CVSS 3.1
          |               |               |
    ┌─────▼───────────────▼───────────────▼─────┐
    │              Hunt Memory                   │
    │  journal · patterns · audit · rate limit   │
    └───────────────────────────────────────────-─┘

Each stage feeds the next. Claude orchestrates everything, or you run any stage independently.

Commands

Core Workflow

Autonomous & Memory (new in v3)

AI Agents

7 specialized agents, each tuned for its role:

What’s New in v3.0.0

The “brain in a jar” is now a bionic hacker.

Vulnerability Coverage

Tools & Architecture

claude-bug-bounty/
├── skills/                     8 skill domains (SKILL.md files)
├── commands/                   13 slash commands
├── agents/                     7 specialized AI agents
├── tools/                      21 Python/shell tools
├── memory/                     Persistent hunt memory system
├── mcp/                        MCP server integrations
│   ├── burp-mcp-client/        Burp Suite proxy
│   └── hackerone-mcp/          HackerOne public API
├── tests/                      129 tests
├── rules/                      Always-active hunting + reporting rules
├── hooks/                      Session start/stop hooks
├── docs/                       Payload arsenal + technique guides
├── web3/                       Smart contract skill chain
├── scripts/                    Shell wrappers
└── wordlists/                  5 wordlists

Installation

Prerequisites

# macOS
brew install go python3 node jq

# Linux (Debian/Ubuntu)
sudo apt install golang python3 nodejs jq

Install

git clone https://github.com/shuvonsec/claude-bug-bounty.git
cd claude-bug-bounty
chmod +x install.sh && ./install.sh

API Keys

export CHAOS_API_KEY="your-key-here"
echo 'export CHAOS_API_KEY="your-key-here"' >> ~/.zshrc

The Golden Rules

These are always active. Non-negotiable.

 1. READ FULL SCOPE        verify every asset before the first request
 2. NO THEORETICAL BUGS    "Can attacker do this RIGHT NOW?" — if no, stop
 3. KILL WEAK FAST         Gate 0 is 30 seconds, saves hours
 4. NEVER OUT-OF-SCOPE     one request = potential ban
 5. 5-MINUTE RULE          nothing after 5 min = move on
 6. RECON ONLY AUTO        manual testing finds unique bugs
 7. IMPACT-FIRST           "worst thing if auth broken?" drives target selection
 8. SIBLING RULE           9 endpoints have auth? check the 10th
 9. A→B SIGNAL             confirming A means B exists nearby — hunt it
10. VALIDATE FIRST         7-Question Gate (15 min) before report (30 min)

The Trilogy

Contributing

PRs welcome. Best contributions:

• New vulnerability scanners or detection modules
• Payload additions to skills/security-arsenal/SKILL.md
• New agent definitions for specific platforms
• Real-world methodology improvements (with evidence from paid reports)
• Platform support (YesWeHack, Synack, HackenProof)

git checkout -b feature/your-contribution
git commit -m "Add: short description"
git push origin feature/your-contribution