The vulnerability stems from improper input sanitization in the click-enlarge feature when caching is enabled. TYPO3's frontend content rendering pipeline (specifically image handling) is the primary suspect. The ImageContentObject::render method is central to generating image markup, and ContentObjectRenderer::getImgResource is a core utility for image processing. Both would interact with user-provided data (e.g., image parameters) that could inject XSS payloads. The lack of escaping before caching would persist the vulnerability. While exact code isn't provided, TYPO3's architecture and the described attack vector strongly implicate these functions.