core practice · system design prompt
Design exercise: Security & Isolation Boundaries
Threat models, least privilege, capabilities, process and VM isolation, side channels, and secure defaults. Implement designOutline() returning non-empty values for: threatModel, trustBoundary, leastPrivilege. Each value must name a concrete mechanism or decision.
The mechanism
Security is control over authority and information flow. Define the adversary, minimize ambient privilege, isolate tenants, validate boundaries, and fail closed. Scope: this card owns the principles — threat modelling, least privilege, capabilities, and failing closed. Which mechanism implements them, and at what cost, is `sandbox-execution-environments`; applying them to an agent is `agent-permissions-sandboxing`.
Expected evidence
A design outline with threatModel, trustBoundary, leastPrivilege plus an explicit failure mode or trade-off.
Try this drill in the practice workspace →
Hints when you need them
- Name the failure mode before selecting the mechanism.