docs: add evaluator guides for remaining released plugins

[olliethedev]

Closes #263

Why

A first-time evaluator should be able to decide whether each released BTST plugin fits before entering installation or API reference material. This extends the approved #262 decision-guide pattern to the remaining released inventory without changing runtime APIs or rewriting the detailed references.

What changed

• adds canonical decision metadata for AI Chat, CMS, UI Builder, Kanban, Comments, Media, Route Docs, and Better Auth UI
• regenerates the versioned packages/cli/plugin-decisions.json artifact with all eleven released records and drift coverage
• applies the shared evaluator header and catalog-card treatment to the eight remaining pages
• keeps exactly one evaluator action and one implementation action on every detail page
• adds authentic generated-app proof for CMS, Kanban, Comments, Media, and Route Docs
• reuses the existing authentic AI Chat and UI Builder proof; intentionally leaves Better Auth UI without a screenshot or demo because neither is verified
• adds deterministic CMS/Kanban product-proof fixtures and preserves Media fixture cleanup
• removes three unsupported Blog RSS claims, replacing them with the implemented metadata/sitemap wording as a narrow factual follow-through

Product facts and boundaries

• UI Builder is client-only and dependent on CMS
• Route Docs is client-only
• Better Auth UI is a client-only companion for an existing Better Auth deployment
• AI Chat names adopter-owned model provider, credentials, billing, access policy, and authenticated persistence
• Comments and Media do not invent standalone demos
• all records use the maintained Next.js 15+, React Router v7, and TanStack Start paths and remain Preview absent an explicit support commitment

Verification

• pnpm lint
• pnpm typecheck
• pnpm test — 195 CLI tests + 1,043 stack tests passed
• pnpm build
• pnpm --filter @btst/codegen plugin-decisions:generate
• pnpm --filter @btst/codegen plugin-decisions:check
• pnpm check:product-proof
• pnpm test:product-proof
• docs TypeScript + production build (56 generated routes)
• playground typecheck + production build
• product-proof capture run twice with byte-identical WebP SHA-256 hashes
• agent-browser: catalog and all eight detail pages; desktop/mobile, light/dark; exactly two detail actions; proof load/alt checks; no evaluator overflow or framework overlay
• scoped axe audit on the evaluator section: 0 violations
• all four playground destinations and Better Auth UI source return HTTP 200

Local browser console output contained only the expected Vercel Web Analytics warning because /_vercel/insights/script.js is unavailable outside Vercel.

Independent review

• Standards: PASS — 0 findings
• Spec: PASS — 0 findings; explicitly accepted the README/docs RSS correction as bounded factual follow-through

This PR remains draft for maintainer and external review.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

2 Skipped Deployments

Project	Deployment	Actions	Updated
better-stack-docs	Ignored		Sep 1, 2026 9:45pm UTC
better-stack-playground	Ignored		Sep 1, 2026 9:45pm UTC

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-01T21:54:54.719745Z	d59413a	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.