About Security: Bug Bounty and Pentesting
Bug bounty and pentest work follows a repeatable shape: confirm the scope, map the target, test one vulnerability class at a time, prove impact and report it. The skills here encode that shape so an agent applies the same discipline on every engagement, with checklists drawn from public reports and named research.
The collection covers the common surfaces. Hunting guides handle authentication bypass, CSRF, CORS, cache poisoning, deserialization, business logic and DOM flaws. Others cover API misconfiguration, CI/CD pipelines, cloud IAM, VPN gateways and Android apps, and the orchestrators route a target to the right set.
When choosing, weigh how strictly a skill frames authorization and scope, because every skill here assumes a program or engagement that permits the testing. Look at how it validates a finding before reporting: a skill that demands browser-proven impact or a second test account saves you from low-value submissions. Check also how it handles evidence, since proof often contains other people’s data.
Use an orchestrator when you start from a bare target and a topic skill when you already know the class of bug. Pair either with a scoring skill so severity is stated the way triagers read it.