Menlo Security Expands MARS to Protect AI Assistants and Coding Agents from Threats
Enterprises have been adopting artificial intelligence agents at a rapid pace, leaving them vulnerable to attacks. Menlo Security is addressing this issue with its expanded Menlo Agent Runtime Security (MARS), a cloud-based platform that protects AI assistants and coding agents from prompt injection and data exfiltration threats.
Mars was introduced earlier this year as a security solution for enterprises deploying AI agents in browser sidebars, like Microsoft Copilot in Edge or Google Gemini in Chrome. The platform sanitizes web pages to prevent attacks before the agent acts, not after the damage is done.
The expansion of Menlo MARS comes at a critical time because more and more companies are integrating AI tools into their operations. According to Gartner, by 2026, 40% of enterprise applications will include task-specific AI agents, up from less than 5% in 2025. This rapid adoption creates new security risks that traditional controls may not be equipped to handle.
Menlo MARS operates at a critical level: attackers aren’t just targeting the model itself but also its surroundings. Attackers target content and integrations that agents rely on, requiring more comprehensive controls. Menlo MARS works this way rather than detecting attacks after they’ve occurred.
The platform offers five key features: isolation and threat removal, file sanitization, adaptive data governance for agents, agent authentication and attribution, and human oversight with forensics. These capabilities enable security teams to approve AI assistants while moving fast without sacrificing safety.
Menlo MARS can run on the same Browser Security Platform that Menlo already uses to govern employee access. This allows security teams to apply a unified policy framework across people and agents, with fine-grained controls specific to each agent type. Its isolation-first design also adheres to emerging standards.
Menlo’s decision to build MARS was driven by customer feedback after 70 conversations in three months. Customers raised concerns about assistants accessing sensitive data, low-code agents inheriting employee credentials, coding agents exposing source code, and autonomous agents exposed online.
The expansion of Menlo MARS is a significant development for companies integrating AI tools into their workflows. By providing comprehensive security solutions for these AI agents, Menlo Security addresses one major concern associated with AI adoption: ensuring the safety of AI assistants in enterprise environments.
Mars protects enterprise AI agents from prompt injection and malware as they browse the web or process files. Its isolation-first design aligns with emerging standards like OWASP guidelines and NIST’s AI Risk Management Framework.
The growing need for security solutions that protect AI tools has led to Menlo MARS’ expansion, focusing on threats such as prompt injection and data exfiltration. By providing comprehensive protection for these agents, Menlo Security addresses a major concern in AI adoption: the safety of AI assistants and coding agents in enterprise settings.
Menlo’s platform supports browser sidebar assistants like Microsoft Copilot and Google Gemini, autonomous agents, AI helpers, and AI coding tools. It can run on the same Browser Security Platform used for employee access governance.
The expansion of Menlo MARS is pivotal as enterprises are increasingly integrating AI tools into their operations. By offering comprehensive security solutions for these AI agents, Menlo addresses a pressing concern in AI adoption: safeguarding AI assistants and coding agents within enterprise environments.
Menlo’s solution supports various types of AI tools commonly used today, including browser sidebar helpers like Microsoft Copilot and Google Gemini. Its design enables running on the same platform as employee access control.
The expansion of Menlo MARS represents a significant milestone for companies relying heavily on AI tools in their operations. By providing comprehensive security solutions specifically tailored to protect these agents, Menlo addresses one major concern tied to AI adoption.