Mission: Field Notes
Field notes on agentic AI security, governance, and the decisions security leaders are making right now. Most pieces are published on the Zenity blog, where I work on agentic AI security; each card links to the original.
Sep 29, 2026
This year's MCP spec changes make the protocol easier to scale and easier to sign into, which is real progress. They do not judge whether an authorized agent is about to do the wrong thing, so teams should plan a runtime layer for that separately.
Read on ZenitySep 23, 2026
Healthcare is connecting AI agents to clinical and payer systems through FHIR and MCP faster than the rules can keep up. The piece argues that seeing what an agent can reach, and stopping actions it was never meant to take, is what lets a pilot grow into an enterprise program.
Read on ZenityWorking through an agentic AI decision? Tell us what you're deciding and what's at stake.
hello@cadre6.com