Series 001 · practical agent safety

Do not give an agent a dangerous tool and call the prompt a fuse.

Agents are useful when they can touch the work. They are dangerous for exactly the same reason. This series is the public-safe version of the discipline: boundaries first, autonomy second.

01

The System Prompt Is Not the Safety System

Why behavioural text is not the same thing as a boundary.

02

Why Useful Agents Need Verifiers

A worker without an independent check is just a confident risk amplifier.

03

The Real Problem With AI Agents Is Context Selection

Most failures start before the model answers: wrong evidence, wrong tools, wrong memory.

04

Why Prompting Is Becoming Spec Writing

A proper prompt is closer to a work order than a magic phrase.

SERIES RULES

What I mean by safe enough to use

NEXT KIT

Start with the checklist. If a workflow cannot pass the checklist, it does not need a bigger model yet. It needs a smaller permission set, a clearer owner and a boring rollback path.

Open all public-safe kits →