Key points
- A ClickFix scam tricks users into pasting commands that steal active session cookies.
- Stolen cookies allow attackers to access Microsoft 365 accounts without passwords or MFA.
- Hidden inbox rules are used to conceal fraud and facilitate business email compromise.
Attackers gain full control of business email accounts by stealing session cookies and manipulating inbox rules, according to Huntress. They need only a user to execute a terminal command, often prompted by a fake AI search result, to initiate the compromise.
In plain English
The attack relies on trust in familiar AI platforms. Users searching for help encounter malicious links disguised as legitimate services. These links lead to pages instructing victims to paste commands into their computer's terminal. This action allows attackers to extract active session cookies from the user's browser.
The background
Huntress researcher Truman Kain described a scenario involving a fictional finance employee named Doris. She clicked a sponsored search result for an AI chatbot and followed instructions to press Command+Space and paste text into Terminal. This ClickFix technique allowed the attacker to retrieve her Microsoft session cookie.
What changes now
With the stolen cookie, the attacker accessed Doris's Outlook without entering a password or multifactor authentication code. They created a hidden inbox rule to archive messages containing specific invoice numbers. The attacker then replied to a treasury department, requesting payment to a new account, and deleted the evidence.
The attacker's path
Teams usually detect this intrusion by monitoring for unusual terminal activity on endpoints. Check whether shell processes are executing encoded network commands shortly after a user interacts with an AI platform. This stage of an intrusion serves to steal active session tokens, bypassing standard authentication controls. Focus on identifying the initial access vector where the victim executes attacker-controlled code on their local machine.
What to do and how to stay safe: ClickFix
- Monitor endpoint logs for shell processes executing encoded network commands, particularly those involving Base64 strings.
- Review Microsoft 365 audit logs for the creation of inbox rules, especially those with minimal or hidden names.
- Educate users to avoid pasting terminal commands suggested by AI platforms or unknown websites.
- Validate payment requests through secondary channels before processing, regardless of the sender's apparent identity.
Step-by-step guide: Phishing vs Spam: How to Tell the Difference and Stop Threats
General security guidance from the RedTeam Log newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How do attackers bypass MFA in these attacks?
They steal active session cookies, which maintain authenticated access without requiring new password or MFA prompts.
What is the role of the AI platform in this attack?
Attackers use fake AI search results to direct victims to malicious pages that prompt terminal command execution.
How long does the compromise typically take?
The entire process, from initial access to financial fraud, can unfold within about 24 hours.
