Connect Clawsmith to your coding agent. Ship products like crazy.Unlimited usage during betaGet API Key →
← Back to dashboard
clawsmith.com/signal/malicious-ai-chrome-extensions-steal-chatgpt-900k-users
IssueUnderservedbrowser_extensionLive

Malicious AI Chrome Extensions Steal 900K Users ChatGPT and DeepSeek Conversations

Two fake AI sidebar Chrome extensions (impersonating AITOPIA) were installed by 900,000 users and silently exfiltrated full ChatGPT and DeepSeek conversation histories to attacker C2 servers every 30 minutes. One carried Google's 'Featured' badge. Both remained live after OX Security reported them Dec 29, 2025. Separately, 287 extensions with 37M users were caught exfiltrating browsing history (HN: 474 pts). The Chrome Web Store's automated scanning missed both campaigns. Demand is emerging for a pre-install extension safety scanner: a 'Carfax for Chrome Extensions' built on HN Feb 2026.

Product Idea from this Signal

A browser extension that detects and blocks other extensions from reading your AI chat sessions in real-time

658

Malicious Chrome extensions steal ChatGPT, Claude, and DeepSeek conversations from users at scale by silently reading DOM content and exfiltrating session tokens. Existing extension scanners only do pre-install static permission checks and miss behavioral exfiltration once an extension is installed. This tool monitors every installed extension's actual runtime behavior on AI chat pages and alerts or blocks when one attempts to read conversation content, scrape tokens, or send data to suspicious domains.

browser securityAI privacyextension monitoringChromeChatGPTdata exfiltration
Competitive172 leadsView Opportunity →

Score Breakdown

HN
658

Gap Assessment

UnderservedExisting solutions leave gaps

ExtensionShield, TrustScan, Extension Auditor exist but are all new/early-stage with minimal users. Google's own vetting demonstrably failed to catch 900K-user malicious extension carrying Featured badge.

Frequently Asked Questions