.jpg)
Zenity Labs has publicly disclosed PleaseFix, a family of critical vulnerabilities impacting agentic browsers, including Perplexity Comet. These flaws enable attackers to perform zero-click hijacking of AI agents, exfiltrate local files, and steal credentials within authenticated user sessions, including interactions with password managers such as 1Password.
Agentic browsers introduce a new computing paradigm where AI agents interpret instructions, maintain authenticated context, and autonomously perform tasks across applications and services. Unlike traditional browsers focused on content display, these systems extend user trust into automated workflows, creating expanded attack surfaces not adequately covered by existing browser or endpoint security controls.
Zenity Labs identified scenarios where attacker-controlled content, such as a malicious calendar invite, triggers autonomous agent execution during routine user requests. This results in zero-click compromises that inherit the full scope of user-authorized access, including local files, tools, and connected services.
“This is not a bug. It is an inherent vulnerability in agentic systems,” said Michael Bargury, co-founder and CTO of Zenity. “Attackers can push untrusted data into AI browsers and hijack the agent itself, inheriting whatever access it has been granted. This is an agent trust failure that exposes data, credentials and workflows in ways existing security controls were never designed to see.”
The PleaseFix family includes PerplexedBrowser, a set of vulnerabilities specific to Perplexity Comet, leveraging indirect prompt injection with two primary exploit paths:
These techniques demonstrate how agentic capabilities amplify traditional attack vectors, shifting from user-targeted social engineering to direct agent manipulation.
Zenity Labs responsibly disclosed the vulnerabilities to Perplexity and 1Password. Perplexity resolved the core browser execution issue prior to public release, while 1Password validated that the vulnerability originates in Perplexity’s agent model.
About Zenity
Zenity is the first security and governance platform purpose built for AI agents spanning SaaS, home grown platforms (Cloud) and end user devices (Endpoint). Trusted by Fortune 500 enterprises, Zenity helps security teams confidently adopt AI by delivering defense in depth with full lifecycle coverage, from agent discovery and posture management to real time detection, inline prevention and response. With an agent centric approach that prioritizes how agents behave, what they access and which tools they invoke, Zenity eliminates blind spots and enforces consistent policy and controls across environments so organizations can innovate with AI without compromising security.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.