Hack The Box is launching AI Range Enterprise Edition to let security teams assess their own AI agents on the roles assigned to them.[1] For businesses, the stakes are concrete: having evidence before delegating a task, then checking that this evidence remains valid as tools evolve.
An enterprise edition to assess defined missions
AI Range has existed since December 2025. The novelty announced on 6 October 2026 is an edition aimed at corporate security teams, with role-based assessment. Hack The Box presents this approach as a way of regularly measuring an agent’s ability to do the work expected of it.[1][3]
A cyber range is a controlled training environment. It allows tools or professionals to be confronted with representative situations without a real incident being the first test. Passing these exercises provides evidence for assessment, the scope of which depends on the scenario and the conditions chosen.
An overall score is not enough to assign responsibility
An agent may recognise suspicious activity without knowing how to build a complete investigation. For a company, defining its mission therefore means specifying the information it can access, the decisions it is allowed to make and the situations that require human validation.
Take an illustrative case: an assistant triages login alerts. The assessment should check the incidents identified, but also the legitimate alerts dismissed and the cases left unanswered. A quick result loses its value if an analyst then has to redo the whole investigation.
Tests must keep pace with changes to the agent
Hack The Box recommends reassessing agents whenever the models, the components that drive their actions, the environments or the threats change.[2] Otherwise, a one-off validation risks describing a configuration that is no longer the one used day to day.
In practice, the team should keep a record of the version tested, the tools available, the permissions and the results. This makes it possible to compare two assessments and determine whether a change genuinely improves the service.
Four criteria for examining the work delegated
This grid is an editorial proposal for preparing an assessment, not Hack The Box’s scoring scheme.
| Criterion | Result to observe | Control question |
|---|---|---|
| Relevance | Incidents detected and false alarms. | Which important cases are missed? |
| Justification | Sources and evidence supporting the analysis. | Can an analyst verify the conclusion? |
| Permissions | Actions carried out within the intended scope. | When must the agent ask for validation? |
| Cooperation | Handing uncertain cases over to a human. | Does the escalation contain the useful information? |
The analyst’s judgement is also put to the test
The announcement also introduces Agentic Operator Competence Scoring, a proprietary system separate from agent assessment. It aims to measure professionals’ ability to understand a recommendation, spot an error and decide when to step in. The AI-augmented SOC Analyst role uses AI-assisted alert analysis and investigation scenarios.[1]
This distinction avoids reducing oversight to systematic approval. A professional must be able to challenge a conclusion, but also to let an authorised operation continue when the available evidence justifies it.
For SMEs, start with a measurable task
An SME could assess an agent on preparing alert files before considering letting it act on its systems. The pilot should compare the quality of the files and the total time spent by the team, including corrections. Pricing, integrations and access conditions for the offering remain to be confirmed with the vendor.
The press release describes a method and announced capabilities. It does not demonstrate a reduction in incidents for every customer. A laboratory result must be complemented by validation in the company’s own context, with limited rights and defined stop criteria.[1]
Professionals become responsible for the quality of delegation
This development gives concrete form to Cybersecurity. Professionals must be able to design a relevant test, examine the traces produced and recognise a poorly supported conclusion. Understanding threats remains necessary to judge the agent’s work.
Their role also involves setting the limits of delegation. They must distinguish a recommendation from an authorised action and organise a handover to humans when information is missing or the risk goes beyond the intended scope.
The value of this skill is then measured in the quality of the decisions made with the tool. Being able to explain why a result is acceptable, or why it must be rejected, becomes part of the defence work.
Trust that must be renewed with use
Our reading of this announcement is that agent assessment could become a regular operational activity. The challenge will be to link each result to a specific mission and a deployment decision. The useful question remains simple: what evidence currently justifies the autonomy granted to this agent?
ALSO READ
These four CCCLX analyses explore further the issues of permissions, protection and system verification.
GitHub Copilot bulks up for business: coding agents come under IT control
VAST Data wants to let AI use sensitive data in a protected environment
Project Glasswing: Anthropic goes on the offensive against software vulnerabilities
ShinyHunters bypasses WAF rules and revives attacks on PeopleSoft
Sources
- 01
Hack The Box (6 October 2026). “Hack The Box Launches AI Range Enterprise Edition to Set a New Standard for AI Agent Competence in Cybersecurity”. Presentation of the enterprise edition and role-based agent assessment.
- 02
Hack The Box. “Cyber Workforce Development for Human & Agentic Teams”. Presentation of the offerings and continuous agent reassessment.
- 03
Hack The Box (3 December 2025). “Hack The Box launches the world’s first AI cyber range to benchmark AI agents and accelerate human-AI teaming across offensive and defensive cyber operations”. Initial announcement of AI Range.
Checked on 6 October 2026: the Hack The Box sources were consulted via their indexed content, as direct access was unavailable. The four CCCLX links were found; the VAST and PeopleSoft pages were checked in their indexed content after direct access failed. The criteria in the table are editorial and do not reproduce the proprietary scoring scheme.
PDFs, presentations, audio: Adobe wants to turn Acrobat into a content factory
Acrobat no longer just wants to open your documents For more than thirty years, Acrobat has mainly been used to read, create, edit, sign and share PDFs. Adobe now wants to give it a much broader role. On 9 September 2026, the company announced new capabilities based...
Salesforce plugs Claude into its CRM: Claudeforce prepares for life after the interface
What if the CRM soon no longer needed to be opened? For decades, using a CRM meant navigating between customer records, dashboards, menus and forms. With Claudeforce, Salesforce and Anthropic want to change this logic: rather than forcing users to go into Salesforce...
Kimi K3, Mistral, DeepSeek: Oracle turns OCI into an AI model hub
Oracle wants to make OCI a one-stop shop for AI Oracle is stepping up its efforts in the AI cloud battle by significantly expanding the catalogue of models available in OCI Enterprise AI. On 4 September 2026, the group announced in particular the integration of Kimi...
Agentic Data Cloud: Google rethinks data for the age of AI agents
AI agents are forcing Google to rethink data architecture AI agents promise to search for information, make decisions and act directly within company systems, but this autonomy depends on one essential condition: accessing the right data and, above all, understanding...
Google Cloud tackles the hidden bill of AI agents
AI agents raise a new problem: how much does a task really cost? As companies move from chatbots to AI agents able to work through multiple steps, the question of cost becomes much more complex. A chatbot generally handles a request and generates an answer, whereas an...
Workday, Navan, Avalara: ChatGPT Work moves into business software
ChatGPT Work is gradually leaving the chat window With GPT-6 Astra, OpenAI is no longer simply trying to improve the quality of ChatGPT’s answers. The company wants to bring its AI closer to the applications in which employees actually work. With the arrival of GPT-6...
From chip to cloud: Groq raises 350 million to scale up
Groq now wants to sell cloud rather than just chips Groq is pursuing a major strategic transformation. Long seen as one of Nvidia’s challengers thanks to its LPUs (Language Processing Units) specialised in artificial intelligence model inference, the group is now...
Muse Code: Meta launches its AI agent to tackle giant codebases
Meta wants to bring AI agents into complex software projects Meta is stepping up its efforts in AI-assisted software development with Muse Code, a new agent designed to work on large codebases, where traditional assistants can quickly lose the context needed to...
Qwen3.8-Max reportedly coded for 16 days without human intervention
Alibaba tests how far an autonomous AI developer can go Alibaba claims to have reached a new milestone in agentic software development with Qwen3.8-Max, its new flagship model launched in early August 2026. In an internal test, the model reportedly worked fully...
Two months, $1.1 billion: River AI is already exciting investors
A spectacular raise for a start-up barely out of the shadows Two months after emerging from stealth mode, River AI has just completed one of the most spectacular funding rounds of the year in artificial intelligence. The start-up founded by Igor Babuschkin, a former...
