Build a threat model and engagement plan prior to execution: define the target, likely adversary profiles, attack paths, success criteria, and safety constraints before any action is taken.
Conduct social engineering, physical, and process-focused assessments to validate the human and procedural layers of defense, not just the technical stack.
Assess emerging technologies — particularly AI/ML systems, LLM applications, and their pipelines — for adversarial, prompt-injection, data-poisoning, model-extraction, and supply-chain exposures.
...
Secure app registrations, enterprise applications, service principals, managed identities, and workload identities, including consent governance and secret/credential hygiene.
Conduct threat modelling and exposure assessments for Entra ID to identify and mitigate identity attack paths (e.g., token theft/replay, consent phishing, illicit grants, privilege escalation).
Design Entra ID tenant recovery and identity incident-response automation to improve recovery time objectives (RTO) and operational readiness.
...