Hack The Box launches AI Range to test AI agents in cybersecurity roles

Hack The Box Ltd. has launched AI Range Enterprise Edition, a platform designed to test the competence of AI agents in cybersecurity roles. The tool aims to address gaps in evaluating AI agents before and after deployment, offering scoring systems for both AI performance and human-AI collaboration.

Product launch focuses on AI agent competence

Hack The Box Ltd. has launched AI Range Enterprise Edition, a platform designed to test whether AI agents can perform cybersecurity tasks effectively in real-world scenarios. According to SiliconANGLE, the tool aims to address a gap in how security teams evaluate AI agents before and after deployment.

The platform introduces a scoring system called Agentic Worker Competence, which assesses AI agents based on their assigned roles. Each agent receives a score and a pass-or-fail mark in simulated environments, with results tracked over time. This allows security teams to determine if an agent is ready for production or requires further refinement.

New testing environments are added regularly to reflect emerging vulnerabilities and attack patterns, ensuring the platform stays current with real-world threats.

Human-AI collaboration under scrutiny

In addition to evaluating AI agents, the platform measures how well human practitioners work alongside them. This is done through Agentic Operator Competence Scoring, which tests whether security staff can identify errors in an AI’s reasoning and know when to intervene.

At launch, the platform supports two AI-augmented roles: penetration tester and security operations center (SOC) analyst. For example, in the SOC analyst role, practitioners rely on AI to review alerts in realistic scenarios, evaluating the AI’s recommendations before taking action.

The scoring system also accounts for situations where practitioners appropriately allow AI-driven work to proceed without unnecessary intervention, ensuring a balanced assessment of human-AI collaboration.

Funding and market reach

Hack The Box raised $55 million in a Series B funding round in 2023, led by The Carlyle Group Inc., with participation from Paladin Capital Group and Marathon Venture Capital. The funding reflects confidence in the company’s approach to cybersecurity training and AI evaluation.

The company serves more than 800 enterprise customers and has built a community of over 4 million cybersecurity professionals. The launch of AI Range Enterprise Edition expands its offerings beyond traditional training to include AI-specific testing and validation.

Context: AI agent infrastructure and adoption

The launch of AI Range Enterprise Edition aligns with broader trends in AI agent infrastructure and enterprise adoption. Companies like IBM and CoreWeave Inc. are collaborating to develop controls for AI workloads, focusing on security, isolation, and identity management. This reflects a growing need for tools that ensure AI agents operate reliably in complex environments.

In parallel, enterprises are integrating AI foundations like TOTVS S.A.’s LYNN, which incorporates governed operational data to enhance AI-driven decision-making. Such developments underscore the importance of testing and validation tools like Hack The Box’s AI Range as businesses deploy AI agents at scale.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

For cybersecurity teams, the launch of AI Range Enterprise Edition fills a critical gap. Most AI agents are deployed based on benchmark tests or vendor promises, but these rarely reflect real-world conditions. Hack The Box’s platform lets operators simulate actual workloads—like penetration testing or SOC analysis—and measure whether an agent can handle them reliably.

The addition of Agentic Operator Competence Scoring is particularly useful. It doesn’t just test the AI; it tests the humans working alongside it, ensuring they know when to intervene or override the agent’s decisions. This dual approach could reduce blind spots in AI adoption, especially as models evolve and drift over time.

For founders and operators, this is a sign of where enterprise AI is headed: less about hype, more about measurable competence. Companies that adopt such tools early may gain an edge in deploying AI agents safely and effectively.

Key takeaways

• AI Range Enterprise Edition is designed to test AI agents in real-world cybersecurity scenarios, not just benchmarks.
• The platform scores AI agents on their assigned roles and tracks performance over time.
• Agentic Operator Competence Scoring evaluates human practitioners working alongside AI agents.
• New vulnerabilities and attack patterns are regularly added to keep testing environments current.
• Hack The Box serves over 800 enterprises and a community of 4 million cybersecurity professionals.
• The company secured $55 million in Series B funding in 2023, led by The Carlyle Group Inc.

FAQ

What is AI Range Enterprise Edition?

AI Range Enterprise Edition is a platform developed by Hack The Box Ltd. to test AI agents in cybersecurity tasks. It scores agents based on their assigned roles and tracks their performance over time.

How does the platform evaluate human practitioners?

The platform uses Agentic Operator Competence Scoring to assess how well human security staff work alongside AI agents. It tests their ability to identify errors in AI reasoning and decide when to intervene.

Which roles does the platform support at launch?

At launch, the platform supports two AI-augmented roles: penetration tester and security operations center (SOC) analyst.

Why is testing AI agents important in cybersecurity?

AI agents may perform well in benchmark tests but struggle in real-world scenarios. Testing ensures they can handle actual workloads and adapt to evolving threats, while also verifying that human operators can work effectively alongside them.

How does Hack The Box’s funding impact this launch?

Hack The Box raised $55 million in Series B funding in 2023, which supports the development of tools like AI Range Enterprise Edition and expands its reach to over 800 enterprise customers.

Related on Lazyfounder

• More ai news
• IBM and CoreWeave co-design workload controls for AI agent infrastructure