Metano, a security startup founded by two cybersecurity veterans, came out of stealth on Nov. 19 with $12 million in seed funding to protect companies from the AI agents they are putting to work

1

. Sierra Ventures led the seed round; Correlation Ventures, High Alpha and 360 One also took part

2

.

Source: SiliconANGLE

Source: SiliconANGLE

Why agents need their own security

Enterprises are deploying AI agents across devices, coding environments, cloud services and SaaS apps, per both reports, and each one adds another place where things can go wrong

1

2

. The company argues existing security tools protect individual layers of infrastructure but lack the visibility to secure agents that run with legitimate credentials and change behavior based on instructions, external content and conditions around them

1

.

"Traditional security tools weren't designed for AI agents that can make decisions and change their behavior as they operate," said co-founder and CEO Neelima Rustagi

1

.

Rustagi says traditional tools were never designed for software that makes its own decisions mid-task, so Metano watches what agents actually do and steps in before autonomy turns into risk

2

.

What the platform does

Metano describes its product as a single layer covering endpoints, cloud infrastructure and SaaS apps, with three parts

1

. Discovery identifies approved and unapproved agents, servers built on the Model Context Protocol, skills and software packages, along with what each can access

2

. A real-time context graph called Hive connects agent tasks, identities, tools and data access, analyzing whole sessions to flag suspicious behavior and judge whether an action matches the agent's intended purpose

1

. Enforcement applies centrally managed policies wherever agents act, including traffic that may skirt traditional network gateways

2

.

The team draws on security backgrounds from ServiceNow, Zscaler, Palo Alto Networks, FireEye and Microsoft

1

. Detections are informed by the company's in-house research arm, MARS Labs, which studies malicious skills, compromised dependencies and vulnerabilities in Model Context Protocol connections

1

. Metano says its detections line up with MITRE's ATLAS catalog of adversarial techniques against AI systems, and with the OWASP Top 10 list for agentic applications

2

.

Sierra Ventures managing partner Mark Fernandes said every security platform shift tends to leave a few companies holding the new control point, and he called agents among the biggest disruptions he has seen

1

.

One customer is already on board: data engineering firm Sigmoid, whose CISO Srivatsa Ramarao says Metano let the company "say yes to AI across the business" while keeping shadow-agent risks down

2

. Per the company, the platform is generally available now, and security teams can be up and running in under 30 minutes

1

. Its free Eval Edition gives 30 days of real-time agent scanning

1

.

© 2026 TheOutpost.AI All rights reserved