Cybersecurity Startups Emerge as AI Agents Gain Autonomy, Leaving Enterprises Vulnerable
A new era of cybersecurity has dawned with the emergence of AIR Security, a startup that’s raised $50 million in funding to tackle the growing threat of AI agents being compromised by malicious actors. As enterprises increasingly rely on AI agents to connect to tools, data, and services, the risk of these autonomous systems being influenced and directed by adversaries has become a pressing concern.
AIR Security’s research has revealed a staggering number of public AI add-ons (over 17,800) that are relying on untrusted external instruction sources. Moreover, AI Skills in the wild have been impersonating companies like Anthropic and OpenAI to bypass security reviews and execute arbitrary code. This has opened up a pathway for data theft, fraud, or unauthorized access while providing little visibility to security teams.
As Yair Saban, co-founder and CEO of AIR Security, puts it: “Every enterprise has a firewall protecting its network. Now they need one protecting their AI agents.” The problem lies in the fact that AI agents are becoming more autonomous, browsing websites, accessing files, and emails on behalf of employees, making them increasingly vulnerable to influence by malicious actors.
The issue is further exacerbated by the growing output from coding agents such as Claude Code, Cursor, Codex, and others. Enterprises have been adopting these tools at an unprecedented pace, but they’re also afraid to deploy them without a seatbelt – and rightfully so, according to AIR Security.
To address this challenge, AIR Security has developed an AI-specific firewall that discovers and evaluates every skill, plugin, MCP server, and add-on across an organization’s AI agent supply chain. This includes screening for external instruction sources, hidden behaviors, and typo-squatted packages masquerading as official developer tools.
If a malicious or vulnerable add-on is detected, security teams can trace every agent and workflow that depends on it – and revoke it across the organization. This process is continuous, ensuring that even if an existing integration is compromised later, trust is automatically revoked.
By providing a marketplace of pre-vetted, certified add-ons, AIR Security offers a safe route to expand agent capabilities without introducing unmanaged risk for its customers. As the threat landscape continues to evolve, it’s clear that AI-specific firewalls like AIR are essential for protecting enterprises from the growing risks associated with autonomous systems.
In practical terms, this means that security teams must start thinking about AI agents in a different way – as an endpoint that requires protection and monitoring. By adopting an AI-specific firewall and taking proactive steps to evaluate and manage their agent supply chain, organizations can mitigate the risk of compromise and ensure the continued integrity of their operations.
Source: SecurityWeek — 2026-09-03