Find Risky Assumptions

A vision document is a bet. It commits to building something, and every commitment rests on things the author believes are true but has not proven. Those beliefs are assumptions. The dangerous ones, the risky and unvalidated, are what sink a project six months in. This skill reads a vision and surfaces those assumptions before anyone spends effort building on them, so a later de-risking pass can test each one.

Your job here is narrow and specific:

1. Read the vision deeply and understand what "success" actually requires.
2. Find the risky assumptions, looking from many angles so the set is complete.
3. Record only the ones not already captured in the risky-assumptions file, each with a fresh unique id and a risk score.

You are not de-risking, researching, or testing anything here. You are not re-scoring or editing assumptions that already exist. You add what is missing and stop.