✨ The Containment Era is here. →Explore ✨

aviatrix logoAviatrix

Overview

Threat intelligence hub

Command Center

Visualize blast radius & threats

Under Active Attack

Executive Summary

In March 2026, Zenity Labs disclosed critical vulnerabilities in Perplexity's AI-powered Comet browser, collectively termed 'PleaseFix.' These flaws allowed attackers to exploit indirect prompt injections, enabling unauthorized access to local files and credential theft without user interaction. By embedding malicious prompts in trusted content, such as calendar invites, attackers could manipulate the AI agent to perform unauthorized actions, including exfiltrating sensitive data and compromising password managers like 1Password. Perplexity addressed these vulnerabilities following responsible disclosure, implementing fixes to prevent autonomous access to local file systems and unauthorized credential manipulation. This incident underscores the inherent security challenges in agentic AI systems, highlighting the need for robust safeguards against prompt injection attacks and the importance of continuous monitoring and updating of AI-driven applications to mitigate emerging threats.

Why This Matters Now

The 'PleaseFix' vulnerabilities in Perplexity's Comet browser highlight the urgent need for enhanced security measures in AI-driven applications. As AI systems become more integrated into daily workflows, the potential for exploitation through prompt injection attacks increases, posing significant risks to data integrity and user privacy. Organizations must prioritize the development and implementation of robust security protocols to safeguard against such vulnerabilities and ensure the safe deployment of AI technologies.

Attack Path Analysis

MITRE ATT&CK® Techniques

Potential Compliance Exposure

Sector Implications

Sources

Frequently Asked Questions

The 'PleaseFix' vulnerabilities are critical flaws in Perplexity's Comet browser that allowed attackers to exploit prompt injection techniques, leading to unauthorized access to local files and credential theft without user interaction.

Cloud Native Security Fabric Mitigations and ControlsCNSF

Aviatrix Zero Trust CNSF is pertinent to this incident as it embeds security directly into the cloud fabric, potentially limiting the attacker's ability to exploit the AI browser and exfiltrate sensitive data.

Initial Compromise

Control: Cloud Native Security Fabric (CNSF)

Mitigation: The attacker's ability to deliver malicious prompts through calendar invites could be constrained, reducing the likelihood of initial compromise.

Privilege Escalation

Control: Zero Trust Segmentation

Mitigation: The attacker's ability to gain unintended access to the local file system could be limited, reducing the scope of privilege escalation.

Lateral Movement

Control: East-West Traffic Security

Mitigation: The attacker's ability to move laterally within the system could be constrained, reducing the risk of accessing additional sensitive files.

Command & Control

Control: Multicloud Visibility & Control

Mitigation: The attacker's ability to establish unauthorized external connections could be limited, reducing the risk of data exfiltration.

Exfiltration

Control: Egress Security & Policy Enforcement

Mitigation: The attacker's ability to exfiltrate sensitive data could be constrained, reducing the impact of the breach.

Impact (Mitigations)

The overall impact of the attack could be reduced, limiting the extent of data breaches and potential exploitation.

Impact at a Glance

Affected Business Functions

  • User Data Management
  • Authentication Services
  • File System Access
Operational Disruption

Estimated downtime: N/A

Financial Impact

Estimated loss: N/A

Data Exposure

Potential exposure of sensitive user data, including emails, calendar events, and stored passwords, due to AI agent hijacking.

Recommended Actions

  • • Implement Zero Trust Segmentation to restrict AI browser access to sensitive resources.
  • • Enforce Egress Security & Policy Enforcement to monitor and control outbound traffic from AI browsers.
  • • Utilize Threat Detection & Anomaly Response systems to identify and respond to unusual AI browser behaviors.
  • • Apply Inline IPS (Suricata) to detect and prevent malicious prompt injections targeting AI browsers.
  • • Regularly update and patch AI browsers to mitigate known vulnerabilities and reduce attack surfaces.

Secure the Paths Between Cloud Workloads

A cloud-native security fabric that enforces Zero Trust across workload communication—reducing attack paths, compliance risk, and operational complexity.

Stop Advanced ThreatsGet a Free Workload Attack Path AssessmentUnder Active Attack?
Cta pattren Image