04 — QUALITY: Case Study Kit
Pass threshold: 90 / 100 on the weighted scorecard, AND all binary gates pass. When to run: After every Mode 1 production, before delivering the three files. Run on each artifact (V1 HTML, V2 HTML, delivery MD) and on the kit's coherence as a set.
Gate 1 — Binary Pre-Build Gate (must pass before scoring begins)
These are blocking. If any fails, halt; do not score the artifacts.
| # | Check | Pass If |
|---|---|---|
| 1 | Profession type set to a valid value | One of: cpa, ea, cfp, ria, insurance, attorney, general |
| 2 | Extraction has zero [NOT SET] values | Every required field has a sourced value |
| 3 | CTA destination URL provided | Non-empty, valid URL |
| 4 | Convertri hosting flag set explicitly | true or false, not blank |
| 5 | Brand source confirmed | Practice owner's design system extracted, OR AOS defaults explicitly chosen |
| 6 | Story count matches extraction content | If extraction has N stories, output has N cards |
| 7 | All numbers reconcile across extraction sections | Savings amount in card header matches Results Grid; no contradiction |
| 8 | Disclaimer template selected | Matches profession type per 02-case-study-terminology.md |
If any of Gate 1 fails: surface the gap, halt production, do not score, do not deliver.
Gate 2 — Convertri Compliance (binary pass/fail, applies when Convertri = true)
Convertri violations break the page on their hosting platform — these are not "lose 3 points" issues; they are blocking.
| # | Check | Pass If |
|---|---|---|
| 9 | Zero var(-- instances in V1 HTML | Every color and design token is hardcoded as a hex value |
| 10 | Zero var(-- instances in V2 HTML | Same |
| 11 | Zero mailto: links in V1 or V2 | Only the configured CTA URL is used |
| 12 | Zero separate |