Keep Reader sitemap HTML-only

[sarthakagrawal927]

What changed

• limits the HTML sitemap to Reader's four canonical public pages: home, FAQ,
  changelog, and login
• adds substantive Markdown for FAQ and changelog so all four pages have
  equivalent agent-readable representations
• aligns the checked-in and Worker-served agent catalogs and llms documents
• keeps six machine resources discoverable through robots, llms, and catalog
  metadata without putting them in the HTML sitemap
• adds focused tests for sitemap, catalog, Markdown, and generated-edge parity

Why

Reader's sitemap mixed two HTML pages with index.md and llms.txt, while the
public FAQ and changelog pages were absent from the catalog and had no Markdown
counterparts. Search and agent discovery could therefore neither interpret the
sitemap as an HTML inventory nor prove route parity.

Exact coverage

• canonical HTML pages: 4/4
• substantive Markdown counterparts: 4/4
• catalog surfaces: 4/4
• machine resources excluded from the HTML sitemap but retained elsewhere: 6

Checks

• pnpm exec vitest run src/agent-surfaces.test.ts — 3 passed
• focused Biome check — passed
• pnpm typecheck
• pnpm cf:build
• pnpm docs:check
• pre-push repository lint — completed with the existing warning-only baseline
• git diff --check

Generated dist churn from the build was restored and is not part of this PR.
No deployment, secret, migration, or production configuration change is
included.

Closes #27

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Updated (UTC)
✅ Deployment successful! View logs	reader	42536c5	Aug 02 2026, 12:23 PM