Prompt injection
Instructions hidden in documents, emails or web pages that the model reads and mistakes for commands. It is the top risk (LLM01) in the OWASP Top 10 for LLM Applications.
Prompt injection, agent permissions, data leakage: we secure LLMs and AI agents before and after release.

An application with an LLM reads text it does not control: documents, emails, web pages, answers from other systems. An agent also has tools and permissions to act. That is why AI security does not stop at traditional controls: you need to check what the model can read, what it can do and what gets recorded.
We use the OWASP Top 10 for LLM Applications 2025 and the OWASP Top 10 for Agentic Applications, published in December 2025, as references.
Instructions hidden in documents, emails or web pages that the model reads and mistakes for commands. It is the top risk (LLM01) in the OWASP Top 10 for LLM Applications.
Personal data, company secrets or the system prompt ending up in answers or in external services.
An agent with broad tools and permissions can carry out actions nobody authorised: the “excessive agency” (LLM06) of the OWASP Top 10 for LLM Applications.
Manipulated training data or components that alter the model’s behaviour. For high-risk systems, the AI Act also requires measures against these attacks.
Model-generated text inserted without checks into code, queries or commands, with effects on downstream systems.
Loops and requests that grow out of control, driving up costs and causing outages.
For high-risk systems, Article 15 of the AI Act requires an appropriate level of accuracy, robustness and cybersecurity, with measures against data and model poisoning, adversarial examples and confidentiality attacks; under the Digital Omnibus these obligations apply from 2 December 2027 for Annex III systems and from 2 August 2028 for those embedded in products. We design the controls following the OISG paradigm, which assigns a metric to every pillar. For how we build agents, see Agentic AI.
It is an attack in which text read by the model (a document, an email, a web page) contains instructions that the model follows as if they came from the user or the developer. With agents, which have tools and permissions, the consequences can be real actions.
No. Filters are heuristic: a useful signal, not a guarantee. That is why we combine them with minimum permissions for every tool, policies on outbound destinations, sandboxes, human approval of actions with impact and an audit trail.
With a perimeter decided in advance: allowed tools, least-privilege permissions, allowed destinations, budgets and exit conditions for the loop. Actions with impact remain proposals that a person approves.
Yes. We assess applications built on commercial models and on open-weight LLMs run on-premise, together with the tools and data they can access.
Fill in the form: your request goes to the dedicated team.