Evaluate the result and decide keep-or-kill

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

You might say…

“My test is done but I'm second-guessing whether the lift is real — I need the decision written down cleanly so I stop reopening it.”

Reads the experiment's outcome against the pre-set success metric and records a keep-or-kill decision with rationale.

Trigger: Use at the close of a single experiment once results are in, to reach the one decision the experiment was designed to produce.

I/O: Experiment spec + outcome data → keep-or-kill decision with rationale

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.