Series

Bounded by Design

The architecture of agentic security, and what the terminology obscures — a six-part series working through a longer paper, one argument at a time.

01 / 06

The Rebrand

The cybersecurity industry didn't build a new kind of system. It found a new word for the old one.

02 / 06

What's Actually Under the Hood

Before you trust anything a security platform's AI tells you, ask what's feeding it.

03 / 06

The Illusion of Reasoning

A platform that can explain why an alert matters looks like it's thinking. Usually, it's tracing a graph correctly.

04 / 06

Two Kinds of Truth

The single distinction the rest of the argument depends on: what's allowed to recommend, and what's allowed to act.

05 / 06

Where It Actually Breaks, and Why Nobody Says So

The failure modes aren't exotic AI failures. They're distributed systems failures wearing a new name.

06 / 06

When the Claim Becomes True

What would actually have to change for the agentic claim to earn its name.

The Full Paper

All six arguments, in their original form, with citations — Bounded by Design: The Architecture of Agentic Security and What the Terminology Obscures, May 2026.

Read the Full Paper (PDF) →