Skill

Scan assumptions and model logic for soundness

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

You might say…

The numbers add up but I'm not sure the model is actually doing what they think it's doing — I need a read on the logic and the assumptions before I sign off on it.

What it does

Reads quantitative work and judges whether the model logic is coherent, segmentation is right, and implicit assumptions are reasonable and surfaced — flagging data gaps and silent assumptions.

Trigger: Use when reviewing data-heavy analysis to catch judgement errors that numerical reconciliation alone would miss.

I/O: Quantitative model or analysis + data inputs → Flagged model logic issues, segmentation concerns, implicit assumptions, and data gaps with suggested surfacing or fixes

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