Skill

Attack the candidate recommendation before committing

Packaged know-how that tells an agent how to do a job well.

You might say…

I was confident in the recommendation. That's exactly when I most need someone to argue the other side.

What it does

Runs the internal adversarial gate on a candidate recommendation — steelmanning the opposite, naming unvalidated assumptions, anticipating the strongest objection — before committing.

Trigger: Use as the final quality check on your own thinking before communicating a recommendation to the client.

I/O: Candidate recommendation, supporting reasoning → adversarial assessment (steelman, unvalidated assumptions, surviving/failing verdict)

Recognise the problem?

The primitives are the commodity part. The fastest next step is a conversation about composing them into something that works for you.

Start a conversation