Ethical Hacking News
Malicious JetBrains plugins have been exposed to steal AI API keys while capturing chatbot conversations with Google Chrome extensions. A coordinated malware campaign has been ongoing since October 2025, posing as AI coding assistants to gain access to users' AI provider API keys. The incident highlights the growing threat landscape and the need for enhanced security measures to safeguard data.
Malicious plugins posing as AI coding assistants were found on JetBrains Marketplace, targeting developers and AI users. The plugins steal user's AI provider API keys and transmit them to an attacker-controlled server. A coordinated campaign has been ongoing since October 2025, with new plugins released as recently as June 10, 2026. Two plugins have garnered over 25,000 downloads each, but their authenticity is disputed due to possible inflation of download counts. Google Chrome ad blocker extensions were found to intercept non-public conversations and model usage data from major AI platforms. The malicious activity is part of a broader trend targeting developer environments through the open-source ecosystem.
THN Exclusive: A growing threat landscape has been exposed by researchers at Aikido Security, highlighting a coordinated malware campaign that targets developers and users of artificial intelligence (AI) services. The malicious plugins, available on the JetBrains Marketplace, pose as AI coding assistants to gain access to users' AI provider API keys.
According to Ilyas Makari, an expert at Aikido Security, "Every plugin poses as an AI coding assistant built on DeepSeek and other large language models, offering chat, commit messages, code review, bug finding, and unit tests." The plugins are designed to function exactly as advertised but steal the user's AI provider API key and transmit it to a server controlled by the attacker.
The campaign has been ongoing since October 2025, with new plugins released as recently as June 10, 2026. Two of the plugins, CodeGPT AI Assistant and DeepSeek AI Assist, have garnered over 25,000 downloads each, although its authenticity is disputed due to possible inflation of download counts.
The malicious activity has also been linked to the capture of chatbot conversations by Google Chrome ad blocker extensions. Researchers at Aikido Security discovered two extensions - Smart Adblocker and Adblock for Browser - that intercept non-public conversations, model usage, and account-tier metadata from major AI platforms such as OpenAI ChatGPT, Anthropic Claude, Gemini, Microsoft Copilot, Perplexity, DeepSeek, xAI Grok, and Meta AI.
The extensions use legitimate public filter lists as a cover while running an undisclosed telemetry channel. The researchers noted that the data collection operation has been codenamed "PromptSnatcher." It's unclear if these practices violate Google's policies for browser extensions.
Aikido Security emphasized the importance of treating plugins as dependencies that run with user privileges and exercising caution when pasting long-lived secrets into unvetted tools.
Furthermore, this malicious activity is part of a broader trend in which threat actors target developer environments through the open-source ecosystem. The researchers highlighted the lucrative nature of LLMjacking schemes, where paid AI services can be resold for illegal purposes.
The findings underscore the need for enhanced security measures to safeguard users' data and protect against such threats.
Related Information:
https://www.ethicalhackingnews.com/articles/Malicious-JetBrains-Plugins-Steal-AI-API-Keys-as-Chrome-Extensions-Capture-Chatbot-Chats-ehn.shtml
https://thehackernews.com/2026/06/malicious-jetbrains-plugins-steal-ai.html
Published: Wed Jun 17 22:38:49 2026 by llama3.2 3B Q4_K_M
