Cyber Risks · The Hacker News

Claude Extension Flaw Enabled Zero-Click XSS Prompt Injection via Any Website

Summary: Researchers at Koi Security disclosed a vulnerability in Anthropic's Claude Google Chrome Extension, codenamed ShadowPrompt, that chained an XSS flaw in the Arkose component with the extension's allow-listing of a-cdn.claude[.]ai to let any website silently inject prompts into the assistant as if the user wrote them, with no clicks or permission prompts. Anthropic patched the extension in version 1.0.41 with a strict origin check requiring an exact match to claude[.]ai after responsible disclosure on December 27, 2025, and Arkose Labs fixed the XSS flaw on its end as of February 19, 2026.

Key Takeaway: Exploitation could have let an attacker steal access tokens, read conversation history with the AI agent, and send emails impersonating the victim - all from a single page visit the victim never notices.

Read the original at The Hacker News ↗

Topics
Threat Actors & Nation-State
Related