Topic
All digests tagged Policy Enforcement
Suggested route
Move from earlier context through the available material to the newest signals. This route is a reading aid, not a difficulty ranking.
Securing AI agents requires moving beyond standing credentials. By implementing OAuth token exchange and using a Security Token Service (STS), organizations can enforce task-scoped, ephemeral access, mitigating the risk of the 'lethal trifecta.'
Read article
NVIDIA OpenShell 0.1 provides a secure runtime boundary for AI agents, preventing compromised agents from executing unauthorized actions. It allows developers to deploy agents in a sandboxed environment where all components inherit consistent policy, credential, and audit controls.
Read article
Learn how middleware extends Managed Deep Agents, enabling custom behaviors like PII redaction and tool call logging. See how to build custom middleware using `wrap_tool_call`.
Read article
OpenShell provides a secure runtime for autonomous AI agents, enforcing isolation and policy through a Control Plane/Gateway architecture. It uses declarative YAML policies and formal verification to prevent agent breakout.
Read article
Instead of giving an agent a token, give it a budget. A budget has four dimensions: how much, how fast, what can be undone, and who notices.
Read article
Jeff Crume explains the AI Security Trilemma: balancing Smartness, Speed, and Security. He details how capability increases attack surface and introduces architectural patterns like AI security proxies to manage risk.
Read article
cMCP is an open-source gateway enforcing policies on MCP servers, creating tamper-evident audit chains. It moves governance beyond simple rules to verifiable trust using confidential enclaves.
Read article