Finding Bad Needles On A Worldwide Scale
summary
Claims that accurate reflected XSS detection is unsolved but improvable; shows a concrete stack: web lab test suit of real and false‑positive cases, a go-based context detect library using HTML5 and JS parsers, and a CD runner called Griffin that orchestrates scanners (scanos, ZAP, PhantomJS, skipfish) in parallel. Demonstrates near‑zero false positives on Yahoo properties, outlines double‑encoding and textarea edge cases, and reports faster, usable dashboards in production.
AI-written summary · about
chapters (26)
- 0:00Start
- 2:20Reflected Server-Side XSS
- 2:57Cross-Site Scripting (XSS)
- 3:20Large Scale Testing
- 4:38Why Webseclab
- 7:12Webseclab Cases
- 8:02Easy Install
- 9:07Webseclab Demo
- 17:04Webseclab and Open Source
- 17:42Arachni & Webseclab
- 19:42Industry Parallels
- 21:32Scanmus internals
- 22:05Scanmus test example
- 22:21Sample payloads
- 23:50Context detections
- 25:41Solution Direction
- 26:27Continuous Delivery
- 28:37CD integration
- 29:47Gaps and Problems (Gryffin)
- 30:21Scanner and Context
- 31:12Parsing "boxes"
- 31:49Injection detection theory (1)
- 34:30Contextdetect Impact
- 35:03Next Steps
- 35:24Lessons
- 41:51Image attributions
related talks
41mNew Methods in Automated XSS Detection
31m360° View of XSS
47mXSS is dead. We just don't get it.
52mDemystifying Front-End Security
40mDifferences Between Web Application Scanning Tools when Scanning for XSS and SQLi
57mThe Last XSS Defense Talk
01h 01mClient-Side Protection Against DOM-Based XSS Done Right (tm)
09m“Debug Thugs”