AI Agent Firewall Startup AIR Security Emerges From Stealth With $50 Million

AIR Security offers a novel AI-specific firewall that evaluates AI skills, plugins, and MCP servers for malicious behavior, excessive permissions, and supply chain risks. It continuously monitors AI add-ons across an organization's AI agent ecosystem, detecting malicious or vulnerable components and enabling security teams to revoke trust and block compromised integrations. This approach addresses risks arising from AI agents autonomously installing tools, connecting to internal systems, and making decisions with limited organizational visibility. AIR Security's research identified widespread use of untrusted AI add-ons and impersonation attempts targeting companies like Anthropic and OpenAI. The firewall also provides a marketplace of pre-vetted add-ons to reduce unmanaged risk.

The emergence of AI agents as autonomous entities within enterprises introduces new attack surfaces, including the risk of malicious or compromised AI add-ons executing arbitrary code, bypassing security reviews, or accessing sensitive data. Without proper controls, adversaries could exploit AI agents to perform unauthorized actions, leading to data theft, fraud, or unauthorized access. AIR Security's firewall mitigates these risks by providing continuous evaluation and revocation capabilities, thereby reducing the likelihood of successful attacks through AI agent supply chains.

This is a newly introduced security product rather than a vulnerability with a patch. Organizations adopting AI agents should consider deploying AI-specific firewalls like AIR Security to continuously evaluate and control AI add-ons and plugins. Since AIR Security is a commercial solution emerging from stealth, no official patches or vendor advisories apply. No direct remediation actions are specified beyond adopting such AI agent firewalls to manage supply chain risks and malicious add-ons.