Hack The Box launches AI range for enterprise security teams

Hack The Box has launched AI Range Enterprise Edition, aimed at enterprise security teams assessing AI agents in cybersecurity roles.

The new edition lets organisations test their own AI agents in role-based cyber scenarios and track results over time. It also introduces AI-augmented roles for penetration testers and security operations centre analysts, alongside a scoring system designed to measure how practitioners work with AI tools.

The launch reflects a wider shift in cybersecurity operations as companies bring AI agents into day-to-day work. That has created a need for ways to assess not only whether an AI system can complete a task, but also whether staff can supervise it, challenge its output and intervene when necessary.

Hack The Box introduced its AI Range platform in December 2025, and says it has since been used by AI labs, government customers and hyperscalers. The enterprise edition extends the platform by letting security teams apply a role-based appraisal process to their own AI agents.

Agent assessment

One part of the offering is what Hack The Box calls Agentic Worker Competence. Under this approach, organisations can assess whether an AI agent can carry out a specific cybersecurity job to the required standard under realistic conditions, using role-based scores, pass-or-fail results for individual environments and records of performance over time.

These appraisals can be repeated as agents, models, software, data sets and operating conditions change. The aim is to help organisations determine whether an agent's performance has improved, declined or remained stable as the surrounding technical environment evolves.

Human oversight

A second part of the launch focuses on human oversight. Through what Hack The Box calls Agentic Operator Competence Scoring, organisations can assess whether cybersecurity professionals have the judgment to review an agent's work, recognise weak reasoning or errors, and decide when intervention is required.

The first AI-augmented roles available are for penetration testers and SOC analysts. In the SOC analyst role, practitioners are placed in scenarios where they use AI to review alerts, investigate suspicious activity and assess recommended actions.

Haris Pylarinos, Founder and Chief Executive Officer of Hack The Box, said the company sees the assessment of human and AI performance as a combined challenge for security leaders.

"As AI takes on more cybersecurity tasks, leaders need confidence in both sides of the workforce," said Haris Pylarinos, Founder and Chief Executive Officer of Hack The Box. "They need to know that agents can do the jobs they are being given, but they also need people with the expertise and judgment to direct that work, verify it and step in when needed. Building cyber readiness means developing human capability and understanding agent capability together."

Many businesses already test whether human staff can perform their assigned roles, while AI systems may be adopted after performing well in a benchmark or a supplier-led test. Hack The Box argues that such checks may not reflect the actual work agents will need to handle on a continuous basis inside an organisation.

That issue is likely to become more important as AI use expands across security teams. Changes to underlying models, software, data inputs and the threat landscape can all alter the way an agent performs, meaning a one-off evaluation may not be enough.

Continuous evaluation

AI Range Enterprise Edition is intended to address that by placing an organisation's own agent in realistic environments and recording how it operates independently. The resulting data can then be used to judge whether the agent is suited to particular work and where human supervision should remain in place.

"Connect the agent, assign it a role and let it work," said Gerasimos Marketos, Chief Product Officer of Hack The Box. "AI Range records how the agent handles the task independently, including where it succeeds and where it struggles. Security leaders can use those results to decide what work an agent is ready to take on and where human oversight is still needed."

Hack The Box also pointed to external recognition in the cybersecurity training and cyber range market. It said it was named a Leader in The Forrester Wave: Cybersecurity Skills and Training Platforms, Q1 2026, and received the highest possible score for Agentic AI Readiness. It added that Gartner named it a Sample Vendor for Cyber Ranges in the Hype Cycle for Security Operations, 2026.

Hack The Box says it supports more than 800 enterprise customers and a global community of more than 4 million cybersecurity professionals.