bug-bounty is the master workflow skill of the Claude-BugHunter bundle. It guides a task from recon and program scope through research, vulnerability testing, validation and reporting, and routes to the topic skills for each bug class and report format.
Its rules tell the agent to read the full scope first, to ask whether the issue causes real harm to a real user and to drop theoretical findings. It also holds checklists for IDOR, SSRF, OAuth, file upload, race conditions, XSS, SQL injection, GraphQL, AI features, cache poisoning, request smuggling and CI/CD. It is intended for assets covered by a bug bounty program you have joined.
What You Get
- Critical rules on scope, impact, time limits and automation
- A cluster-hunting method that turns one bug into a chain, with known chains in a table
- Checklists for the common web, API, mobile, AI and CI/CD bug classes
- A seven-question gate, four pre-submission gates and a CVSS quick guide
- Routing notes to the methodology, recon, triage and report-writing skills
Common Use Cases
- Starting a bug bounty engagement and wanting a single skill that routes to the rest
- Checking whether a finding clears the gate before spending time on a report
- Looking up how to escalate a low-severity bug into a reportable chain
What Makes It Unique
It is the orchestrator of the bundle: it keeps the whole pipeline in one place and then hands off to specific skills by topic. It also carries operator notes on when to use the orchestrator, when to go straight to a class skill and why loading every hunting skill at once wastes context.
Under The Hood
The skill is one long SKILL.md of instructions with no bundled scripts. Following it, the agent runs shell commands for recon and testing such as subdomain discovery, URL collection, fuzzing and secret scanning, and it sends requests to the target hosts and to public sources. It writes output files and notes to the working folder, clones repositories with git, and its install notes fetch tools from the network. It also covers credentials: it shows patterns for finding leaked keys and tokens and tells the agent to prove what a found key can reach. One section on CI/CD workflows quotes prompt-injection phrases as examples of what to look for. Run it only against assets you are authorized to test.