Assertion Labs / Advisory
Understanding the authority of AI systems.
Examine how AI applications access data, exercise authority and interact with other systems. Evaluate the architecture and the evidence behind its security claims.
Discuss your work →Approach
Examine the boundaries. Make the evidence clear.
Each review begins with the system, its intended use and the decision it needs to support. Scope, methods and deliverables are agreed together before work begins.
Read an illustrative assessment →Context
Engineering and research teams evaluating the security of AI-enabled software.
Preparation & access
An architecture diagram, a workflow walkthrough and selected authorized code or configuration. Agree on a secure transfer channel after scoping; do not upload confidential materials through the inquiry form.
Work & deliverables
- An agreed system boundary and evaluation question
- An annotated architecture and trust-boundary map
- Prioritized findings linked to reviewed evidence and explicit limitations
- A technical assessment and discussion of the findings
Scope & limitations
Scope, access and methods are agreed for each engagement. Penetration testing, production access and implementation require separate authorization. A review is not a certification or a guarantee that a system is secure.
Offered subject to fit, conflict review and availability. A conversation is not acceptance of an engagement.
How we work
Begin with a bounded question.
- Define the decision, boundaries, conflicts and available evidence.
- Agree on the work, deliverables, access and review process.
- Examine claims and record findings with their limitations.
- Discuss implications and the next decision.