Proof

Overview

Most of this site makes a case: tribal knowledge is a bottleneck, contracts fix it, and the fix compounds. A case is only worth as much as the mechanics underneath it, so this section is where those mechanics get shown rather than asserted. How does a legibility score actually get computed. Why does extraction beat a rewrite on the math, not just on vibes. What happens when you go look at the toolkit's source instead of taking a claim about it on faith.

None of the three pages below ask you to trust a metaphor. They walk through the scoring rules, the extraction priority logic, and the fact that the CLI and its schemas are sitting in the open where anyone can read them.

Pages in this area

• Legibility Scoring Methodology

  How four questions and a half-point scale turn "this code feels risky" into an actual number.

• The Case for Extraction Over Rewrite

  Why a rewrite spends senior engineering time on cleaner code that still doesn't declare what it knows.

• Open Source Toolkit

  The CLI, the schemas, and the validation logic are public, so the methodology isn't a marketing claim.