Open to principal-scope roles · Phoenix, Arizona · Remote or hybrid
Learn AI · On the map · Agents · idea 32 of 44
Checks in code that run before or after the model acts, so certain things can never happen no matter what the model says.

In one line
Checks in code that run before or after the model acts, so certain things can never happen no matter what the model says.
Picture it
A locked gate, not a sign that says please do not enter.
Under the hood
Pre- and post-tool hooks, schema validation, allow-lists, rate limits, kill switches. A hook is a rule only when it is a check in code that can block the call; some hooks only log, and some ask another model. When money or safety is on the line, enforce in code at the boundary the action crosses.
Builds on
On a guided path
Path 5: Running it safely, stop 2 of 11.
Marking "I get this" is self-reported and stays in this browser. The same mark shows on the map.