bug-bounty

Guides end-to-end bug bounty workflows from recon through validated vulnerability reporting.

1|Updated Aug 21, 2026
One-click install
npx skills add https://github.com/marcboggs/BMAD-AppSec-Orchestrator --skill bug-bounty-marcboggs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: bug-bounty
Source: https://github.com/marcboggs/BMAD-AppSec-Orchestrator/tree/main/.claude/skills/bug-bounty
Command: npx skills add https://github.com/marcboggs/BMAD-AppSec-Orchestrator --skill bug-bounty-marcboggs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Bug bounty hunting involves many disconnected phases — recon, learning the target, hunting dozens of vulnerability classes, validating exploitability, and writing reports that programs accept. This Skill consolidates the entire pipeline into one workflow with checklists, bypass tables, and validation gates so hunters avoid wasted effort on theoretical or out-of-scope findings. ## Core Features & Use Cases - Full Recon Pipeline: Subdomain enumeration, live host probing, URL collection, nuclei scanning, JS secret extraction, cloud asset enumeration, and HackerOne scope retrieval with ready-to-run commands. - Vulnerability Hunting Checklists: Deep testing guides for IDOR, SSRF, XSS, SQLi, OAuth/OIDC, GraphQL, race conditions, file upload, business logic, and LLM/AI agent attacks (ASI01-ASI10), including bypass tables for SSRF IPs, open redirects, and file uploads. - A-to-B Bug Chaining: A cluster-hunt protocol that escalates single findings (e.g., SSRF to cloud metadata to IAM credential theft) into higher-payout chains. - Validation & Reporting: A 7-Question Gate, impact-first triage rules, CVSS 3.1 scoring, and report templates to kill weak findings before submission. - Use Case: Starting on a new HackerOne program, pull the program scope, run the standard recon pipeline, fingerprint the tech stack, then hunt IDOR variants across API endpoints using the two-account testing method. ## Quick Start Ask the agent to start a bug bounty recon and hunting session against a specified in-scope target domain.

Frequently Asked Questions about bug-bounty

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I start bug bounty recon on a new target?▼

Run the standard recon pipeline: enumerate subdomains with subfinder and assetfinder, resolve and probe live hosts with dnsx and httpx, collect URLs via katana, waybackurls, and gau, then scan with nuclei. Finish by extracting secrets from JavaScript files and checking quick wins like exposed .git directories and S3 buckets.

How do I test for IDOR vulnerabilities in an API?▼

Create two accounts (attacker and victim), capture all object IDs in requests as the victim, then replay those requests with the attacker's session. Test all ten variants including body parameters, batch endpoints, nested paths, method swapping, version rollback, and header injection like X-User-ID.

What tools are needed for bug bounty hunting with this workflow?▼

Core Go tools include subfinder, httpx, dnsx, nuclei, katana, ffuf, dalfox, and interactsh-client. Optional installs cover arjun for parameter discovery, trufflehog and gitleaks for secret scanning, sqlmap for SQL injection, and semgrep for static source code analysis.

How do I bypass SSRF protections to reach cloud metadata?▼

The workflow provides eleven bypass techniques: decimal, hex, octal, and short IP notation, IPv6 and IPv4-mapped addresses, redirect chains, DNS rebinding, URL encoding confusion, unicode numerals, and protocol smuggling via gopher. Target 169.254.169.254 for cloud metadata and internal services like Redis on port 6379.

When should I not submit a vulnerability report?▼

Do not submit theoretical bugs that require multiple preconditions, dead code, source maps without secrets, DNS-only SSRF callbacks, or open redirects without an account takeover chain. The 7-Question Gate requires demonstrating real harm — stolen funds, leaked PII, account takeover, or code execution — against a real user right now.

Does this workflow cover LLM and AI agent security testing?▼

Yes, it covers prompt injection, indirect injection via processed documents, chatbot conversation IDOR, system prompt extraction, RCE through code execution tools, and ASCII smuggling. It also maps the OWASP ASI01-ASI10 attack classes for agents with tool access.