Description

A security researcher has recently introduced a novel exploitation called “SVG clickjacking”, dramatically upgrading classical clickjacking into a fully interactive, logic-driven threat. Instead of simply tricking users into clicking an invisible button, this method lets attackers overlay a fake user interface (UI) that can respond to — and even read — the real state of a victim site, guiding users through complex tasks like typing text, toggling boxes, or filling forms — all while the user believes they are interacting with a harmless overlay. At the heart of SVG clickjacking is the misuse of standard SVG filters (such as feColorMatrix, feDisplacementMap, feComposite). What are normally visual-effect tools become logic-gates (AND, OR, XOR) embedded in the browser’s rendering pipeline. These filters can sample rendered pixels from a cross-origin iframe (i.e. the target site embedded inside another page), effectively “observing” whether certain UI elements — dialogs, pop-ups, checkboxes, error messages — are visible or active. Based on that visual feedback, the attacker’s overlay updates in real time, steering the user’s interaction flow. In a proof-of-concept, the researcher tricked a user into creating a document and typing data in a site like Google Docs by having them believe they were solving a simple CAPTCHA or puzzle. The implications are serious: by converting pure rendering tools into a dynamic logic engine, this technique bypasses many of the classical assumptions about web-UI security. Traditional defenses — like preventing framing via X-Frame-Options or Content-Security-Policy frame-ancestors, or treating clickjacking as a static overlay risk — may no longer suffice. As this method enables not just click-tricking but reading UI state and adapting overlay behavior accordingly, attackers could perform complex unwanted operations, including data entry, toggling settings, or potentially exfiltrating information without ever injecting code into the target site.