Defense in Depth

Layered security strategies that use multiple overlapping safeguards—isolation, least privilege, sandboxing, verification—so that no single failure compromises the whole system.

Reading List

Agentic Systems

Capping the Blast Radius: Engineering Secure AI Agent Containment

Jun 4, 2026223

Effective AI agent security requires capping the potential 'blast radius' through deterministic environmental containment rather than relying on probabilistic model safeguards or human oversight.

AI Agents Sandboxing Anthropic Defense in Depth AI Safety

Under the Hood

Engineer AI for Failure: Contain Prompt Injection

Sep 26, 2025115

Stop prompt-injection harm by engineering AI like machines: assume failure, isolate, constrain, and verify.

Prompt Injection AI Safety Sandboxing Defense in Depth