Attack the candidate recommendation before committing
Runs the internal adversarial gate on a candidate recommendation — steelmanning the opposite, naming unvalidated assumptions, anticipating the strongest objection — before committing.
In / outCandidate recommendation, supporting reasoning → adversarial assessment (steelman, unvalidated assumptions, surviving/failing verdict)
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
Single-judgement skill that runs the internal adversarial gate: takes the candidate recommendation and actively argues it is wrong — steelmanning the opposite position, naming the assumptions not yet validated, anticipating the client's strongest objection, and judging whether the recommendation survives. Ends with a settled, defensible position or a deliberate decision to withhold pending more information. Used as the final quality check on the advisor's own thinking before they commit.
Trigger: Use as the final quality check on your own thinking before communicating a recommendation to the client.
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