Unverified50% confidenceFactExact time
In the second link of the attack chain, the attacker injects a new iframe without the sandbox attribute via innerHTML, thereby bypassing the original sandbox restriction and controlling an unrestricted browsing context.
1
Sources
50%
Confidence
Long-term
Relevance
11/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe first link in the attack chain exploited the postMessage entry point of the MathLive virtual keyboard, a component that renders content using innerHTML and does not restrict the message origin, allowing untrusted code inside the sandbox to inject custom HTML into the parent page.75% similarUnverifiedThe frontend connects directly to the sandbox via WebSocket, for viewing the code editing interface in Code Server and watching the real-time screen of the browser inside the sandbox.63% similarUnverifiedThe LaTeX renderer uses an <iframe sandbox> mechanism to defend against XSS risks, layered with DOMPurify for HTML sanitization.60% similarUnverifiedAt the application layer, scroll anchoring can be implemented by recording the getBoundingClientRect offset of the target element before and after content insertion and manually calling scrollBy for correction, which is suitable for scenarios such as simultaneous content injection at both ends that native behavior cannot cover.60% similarUnverifiedIn production environments, browser automation is usually used together with sandbox isolation, operation replay, and human fallback mechanisms, because it faces engineering challenges such as page structure changes, asynchronous loading, login state and anti-scraping mechanisms, and real side effects.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1121900API
curl https://kongchang.com/api/v1/knowledge/claims/1121900MCP
get_claim(id=1121900)