3 min read
|
Saved February 14, 2026
|
Copied!
Do you care about this?
SquareX's research reveals a concealed MCP API in the Comet browser that allows extensions to execute commands and access user devices without permission. This breach of security principles raises concerns about potential exploitation and the lack of transparency surrounding embedded extensions.
If you do, here's more
SquareX uncovered a serious security issue in the Comet browser, revealing a hidden API known as MCP (chrome.perplexity.mcp.addStdioServer). This API allows extensions to execute local commands on users' devices, a capability that traditional browsers restrict. The lack of clear documentation around the MCP API raises alarms, as users remain unaware of the potential risks. Kabilan Sakthivel from SquareX pointed out that Comet has bypassed long-established security measures, putting user trust at risk.
The MCP API is currently utilized by the Agentic extension, which can be activated through the perplexity.ai page. While thereβs no evidence of current misuse, the potential for exploitation is high. A single vulnerability or phishing attack could give malicious actors unprecedented access to user devices. SquareX demonstrated this risk by showing how a disguised malicious extension could utilize the MCP API to execute harmful software like WannaCry.
Moreover, the extended functionality of these embedded extensions is obscured from users, making it impossible for them to disable them even if they suspect a compromise. This lack of transparency means users are blind to significant security risks. Other AI browsers may have similar extensions, but the MCP API appears unique to Comet for now. SquareX urges AI browser developers to improve documentation, undergo security audits, and provide users with the ability to disable risky features. The push for accountability is critical to prevent a dangerous trend in browser security.
Questions about this article
No questions yet.