About

SAFE SQUARE is a security technology company. We research and develop technologies for Sovereign Security AI and put them to work through professional security services.

Sovereign Security AI is the capability we are working towards, not a single model or a released product.

Our research spans AI models and agents, exposure intelligence, and offensive and defensive techniques. We also study how professionals can analyse more deeply, make better-informed judgements, and work alongside multiple people and AI agents in a shared environment.

We do more than collect knowledge and experience. We develop frameworks and systems that connect collection, processing and management with delivery and use, so people, agents and models can apply the knowledge they need in the form that works for them.

We use capable external AI technologies while also improving models for security work and developing local models that are easier to deploy, customise and operate. Choosing or combining models is one of many approaches. Our goal is to improve knowledge, tools, verification and collaboration together, reducing the effect that differences between models and environments have on the quality of work.

We apply our research and practical experience to exposure intelligence, vulnerability assessment and penetration testing, and security capability development for people and AI. We also offer comprehensive security health checks, annual cybersecurity care, custom engineering and ongoing collaboration with client teams. Every service can be contracted on its own or combined with others.

Research and development at SAFE SQUARE centres on two core organisations of its own: tty.link and Return-Oriented AI Lab.

People

• Security assessment and validation

  Analysis and validation from an attacker's perspective and threat investigation, with findings and real impact explained to decision-makers and engineers.

• Education, training and evaluation

  Turning knowledge and experience from research and engagements into learning, training and evaluation, so that the security capabilities of people, AI and teams keep developing; designing courses and lab environments, and running hacking competitions and drills.

• Business and operations

  Understanding each client's challenge and matching it with the right people and realistic terms; running engagements to the agreed schedule and scope, reviewing the quality of what we deliver, and protecting the information entrusted to us as agreed.

• Co-Intelligence Security

  A way of working in which multiple professionals and AI agents investigate, analyse, develop and verify together. People contribute expertise throughout the process, from defining the problem to improving the result—not only at the final approval stage. See how the work is run, controlled and verified

• Agreed in writing

  Before we start, we agree the goals, targets, scope and schedule in writing.

• Only what is approved

  We do only the work that has been approved, within the agreed scope and authorisation. AI and tools run only inside that same scope.

• Facts kept apart from assumptions

  Results separate what we confirmed, what still needs a closer look, and what this scope did not cover.

• Protecting your information

  We ask only for the information we need, receive it through approved channels and handle it within what was agreed; how it is kept and disposed of is settled before we start. We keep your private information separate from public observations, and reuse project material as knowledge or for model training only where permission and the agreed scope allow.

• Remote and on-site

  We work remotely or on site, depending on your environment, schedule and the nature of the work. Team members living in Singapore, Korea, Japan, Indonesia and Germany can join you on site when needed.

• Working languages

  We consult in Korean, English, Japanese, Indonesian and German, and by email in other languages too.