[codex] Upgrade full product evaluation loop

[mberman84]

What changed

• upgrades the existing full product evaluation loop into a production-grade, exhaustive QA workflow
• requires a sanitized production-scale local dataset and production-like settings
• inventories every user-facing surface with documented acceptance criteria and finite risk-based edge cases
• records reproducible bugs, reviews shared causes, fixes holistically, and reruns the full inventory
• regenerates the public catalog, loop page, feed, sitemap, and skill fallback

Why

The prior loop sampled major capabilities but did not require production-like local data, exhaustive interactive-surface coverage, or a whole-system review of discovered bugs.

Validation

• node scripts/check.mjs
• npm test --prefix worker (15 tests)
• local browser render of the generated loop page
• autoreview clean after addressing its acceptance-gate and bounded-edge-case findings