# Changelog

## 0.1.0 - 2026-09-28

First public release under MIT.

- `rank()` orders capabilities by readiness gained per unit of effort,
  `(100 - score) × weight ÷ effort`, not by score.
- Unassessed capabilities (`score: null`) are listed separately and never
  ranked.
- `effort: 'unknown'` weighs the same as `medium`.
- Every entry carries its binding constraint and next action; an entry with
  neither is reported as "a position, not advice".
- `rankingTable()` prints in rank order; `toReport()` and `formatRanking()`
  produce an evidence-record report.