Agentic offensive security platform provider Terra Security Inc. today announced the launch of continuous exploitation validation for network infrastructure, extending the company’s platform beyond web applications and artificial intelligence systems.

Terra’s platform uses swarms of hundreds of AI agents paired with human reviewers to probe customer environments for exploitable weaknesses. With the addition of network coverage, security teams can validate vulnerabilities across web apps, AI systems and infrastructure from one console. Findings are ranked by real exploitability and business impact rather than raw severity scores.

The release seeks to address the issue whereby traditional penetration testing has treated each surface in isolation, with one vendor scanning the network, another testing web applications and AI systems often going untested altogether. The result, Terra says, is fragmented reporting and slow remediation when adversaries are increasingly automating reconnaissance and lateral movement using AI tooling of their own.

Network findings now appear alongside web and AI vulnerabilities in a single connected view, with consistent audit trails for compliance teams. Terra claims that the unified approach can replace multiple point tools and surface multi-vector attack chains that siloed scanners miss.

“When we founded Terra, we believed that continuous, agentic offensive security across the full attack surface was a necessity,” co-founder and Chief Executive Shahar Peled said in a statement. “AI-powered adversaries don’t test one layer at a time, and neither should we.”

Peled added the expanded platform now covers web and internal applications, AI systems and network infrastructure under the same agentic model with human-in-the-loop oversight. Terra is selling the consolidated platform to chief information security officers who currently run red-teaming, application testing and vulnerability management through separate vendors.

Generative AI adoption has made the pitch easier, with each new model integration and third-party application programming interface adding to the external attack surface enterprises need to test. Terra says its agents verify findings for real exploitability before alerting security teams, with auto-remediation built in for issues that can be fixed without human intervention.

The network capability is in public preview for existing Terra Platform customers, with general availability not yet announced.

Peled, along with Daniel Bernard, chief business officer of CrowdStrike Holdings Inc., and CJ Moses, chief information security officer and vice president of security engineering at Amazon Web Services Inc., spoke with theCUBE, SiliconANGLE Media’s livestream studio, in September. He discussed how the company uses offensive security to mimic real adversaries and surface what is actually exploitable rather than merely vulnerable.

Image: Terra Security