Glemad

    Research

    Model research

    We study how a model keeps an accurate picture of changing infrastructure, reasons about what an attacker may be trying to do, and acts only within explicit policy, with the evidence behind each decision kept on record.

    The questions that guide the work

    Live infrastructure state

    How can a model maintain a useful account of an environment that changes while it is being observed?

    Attacker intent

    What evidence separates an attack path from an unusual but legitimate change?

    Cross-domain reasoning

    How should evidence from different infrastructure domains change the same threat hypothesis?

    Uncertainty

    How should the system behave when evidence is incomplete, contradictory, or late?

    Policy boundaries

    Can authority be represented clearly enough to govern a response while an incident is unfolding?

    Controlled response

    What is the smallest defensible action that can interrupt an attack path?

    Evidence preservation

    Can another person reconstruct why the system reached a conclusion and what followed?

    Evaluation

    How do we test reasoning quality when the safest answer may be to wait, escalate, or refuse an action?

    Capability does not grant authority

    A model can identify a plausible response without being permitted to carry it out. Policy decides what can be proposed, what needs approval, what can proceed and when the system must stop, and that check is part of the reasoning record.

    What a model can identify contains what policy permits, which contains what proceeds.

    From observation to review

    State

    What is happening across the environment now?

    Intent

    Which explanation best accounts for the evidence?

    Authority

    What is permitted under current policy?

    Response

    What controlled action is proportionate?

    Evidence

    What must remain available for review?

    Read the research as it develops

    Research index