AI-Powered Bug Hunting
summary
Claims $1M since 2022 and about 900 valid submissions while demonstrating concrete workflows: asset discovery with amass producing ~27,188 subdomains for nasa.gov, parsing chunks via llm (simonw) to find GitLab/Git instances, exploiting misconfigured GitLab APIs to dump projects, using shortscan to enumerate IIS short-name files and extract DLLs containing API keys, then using kaido to authenticate and escalate to remote upload/aspx shells, plus automated username enumeration and Zip Slip POCs that overwrite web app files.
AI-written summary · about
chapters (14)
- 0:00Intro
- 2:12What's a bug bounty?
- 3:15$1M since 2022
- 3:52Easier with AI
- 6:09Applied AI for bug bounties
- 6:33Asset discovery
- 8:30Hacking NASA
- 14:03Insecure direct object reference
- 15:46Unauthenticated access to the API leaks user PIl
- 19:40IIS short name enumeration
- 26:38In collaboration with Shubs & Rens
- 32:08Demo
- 34:13Final thoughts
- 34:41Outro
related talks
32mAutomating Web Application Bug Hunting
46mThe Bug Hunter's Methodology 2.0
38mHackers Gonna Hack: Identifying and Fixing Vulnerabilities Proactively
35mBug Bounty Hunting
38mHacking Demos, Dirty Secrets, Dangerous Lies, Asset Intelligence
29mFlagship Projects: Amass
42mHidden Chains: Revealing High-Impact Bugs from Bounty submissions
37mAI-Powered Slopsquatting: Is Your Software Supply Chain Compromised?
