Coordinated Chrome Extension Campaign Hijacks Active ChatGPT Sessions
Security researchers have uncovered a coordinated campaign abusing Chrome browser extensions presented as ChatGPT productivity and enhancement tools. Unlike isolated malicious add-ons, this operation involves at least 16 separate extensions, all sharing nearly identical code structures and delivery logic, indicating centralized development and intent.
The extensions were distributed through the official Chrome Web Store and positioned as tools designed to improve productivity, automate prompts, or enhance ChatGPT workflows. Their branding and descriptions closely mirrored legitimate AI assistant extensions, significantly lowering user suspicion during installation.
How the attack works
Once installed, the extensions inject code directly into pages where ChatGPT is accessed. Rather than exploiting the platform itself, the attackers target the browser environment and active user sessions.
The malicious logic hooks into browser-level request handling, allowing the extensions to monitor outbound traffic from ChatGPT pages. When authentication headers or session tokens are detected, these values are silently extracted and transmitted to attacker-controlled infrastructure.
Because session tokens represent an already authenticated user, possession of these tokens allows attackers to access accounts without passwords, MFA prompts, or triggering login alerts. In practice, this provides full access to conversation history, saved context, and any connected services accessible through the compromised session.
Why this matters
This campaign represents a shift from traditional credential theft toward session-level compromise, where attackers bypass authentication entirely by operating inside trusted browser contexts.
The coordinated nature of the operation, multiple extensions built on the same malicious framework, suggests deliberate scaling rather than opportunistic abuse. While the number of confirmed installations remains relatively limited, researchers note that adoption could grow rapidly as AI-focused browser tools become more common in both personal and corporate environments.
DIAMATIX Perspective
What makes this campaign particularly relevant is that nothing “breaks” in the classic security sense. There is no vulnerability exploitation, no malware dropper, and no suspicious login activity.
Instead, trust is abused at the browser extension layer. Once an extension is granted permission to operate, it effectively becomes part of the user’s trusted workflow. In environments where AI tools are increasingly embedded into daily work, this blurs the boundary between assistance and execution.
Defensive focus should shift toward:
stricter extension governance and allowlisting
monitoring of browser-level behavior, not just endpoints
treating AI-related extensions as privileged software, not convenience tools
Session integrity is now as critical as credential protection.
Related reading:
AI-driven attacks are increasingly targeting workflows rather than infrastructure. We explore this shift in more detail in our LLM Security 101 series.
LLM SECURITY 101 — PART 1: Understanding the Basics & Key Early Risks
LLM SECURITY 101 — PART 2: Advanced Risks and Practical Safeguards for Everyday AI Use
Sources
LayerX Research. Analysis of coordinated Chrome extension campaigns targeting ChatGPT sessions
Google Chrome Extension Security Documentation
Independent security research on browser-based session hijacking techniques
