Make readiness dashboard audit-honest

[dorkalev]

Summary

Replace the misleading all-green dimension average with a transparent weighted in-scope criterion evidence maturity metric.

Tickets

Fixes #18

Changes

• compute the primary gauge from in-scope criterion weights and evidence credit
• separate pass/fail/unknown/n-a observation counts from criterion status
• mark Type I and Type II as not determined and remove audit-outcome language
• record commit, workflow run, generator, and JSON signature status
• reset incompatible legacy gauge history and document the rubric
• add regression tests for the metric, n/a denominator, and audit-opinion disclaimer

Test Plan

• cargo test --manifest-path actions/shadow-ci/Cargo.toml --locked
• cargo test --manifest-path website/app/Cargo.toml --locked
• render the prior 100% report and confirm 61.2% maturity, 14/36 verified, 8 not started, 39/39 applicable observations, and 1 n/a