Goal: turn the standalone public explainer page(s) on GitHub Pages into one small, coherent "How MateMate Works" site, borrowing the information architecture of Google's "How Search Works" — question-based section headers, progressive disclosure (overview → deep dive), consistent modular cards, and a light shared top-nav across pages.
Current state (main @ cf20ef7, 2026-07-12)
docs/architecture/index.html — live at testtest126.github.io/chess/architecture/ ("What is MateMate?"). Self-contained, theme-aware, interactive.docs/agent-interaction-guide.md — Markdown source describing how the concurrent Claude sessions coordinate. This is the content source for the "How it was built" page.- No
docs/index.html root hub exists (testtest126.github.io/chess/ has no landing page).
- The "How the Agents Talked" page at
docs/the-fleet/index.html does not exist yet despite being referenced as prepared — it needs to be authored.
Scope
- Author
docs/the-fleet/index.html — a self-contained, theme-aware "How the Agents Talked" page (how MateMate was built) derived from docs/agent-interaction-guide.md, matching the architecture page's design system.
- Add
docs/index.html — a "How MateMate Works" hub with plain-language, question-based entry points: What is MateMate? → architecture, How was it built? → the-fleet.
- Add a lightweight shared top-nav tying all three pages together with consistent card styling and overview-to-detail progressive disclosure.
Constraints
- Each page stays a single self-contained HTML file — no external fonts/scripts/assets (Pages renders as-is).
- Preserve the existing theme-aware (light/dark +
data-theme) design and reduced-motion handling.
- PII-clean: no emails, surnames, bundle ids, or private paths; scan before publishing.
Done looks like
- Three pages under
docs/ cross-linked by a shared nav; hub fronts two question-based entry points; both deep pages self-contained and theme-aware; PII scan clean. GitHub Pages publishes on merge to main under /docs.
Goal: turn the standalone public explainer page(s) on GitHub Pages into one small, coherent "How MateMate Works" site, borrowing the information architecture of Google's "How Search Works" — question-based section headers, progressive disclosure (overview → deep dive), consistent modular cards, and a light shared top-nav across pages.
Current state (main @ cf20ef7, 2026-07-12)
docs/architecture/index.html— live attesttest126.github.io/chess/architecture/("What is MateMate?"). Self-contained, theme-aware, interactive.docs/agent-interaction-guide.md— Markdown source describing how the concurrent Claude sessions coordinate. This is the content source for the "How it was built" page.docs/index.htmlroot hub exists (testtest126.github.io/chess/has no landing page).docs/the-fleet/index.htmldoes not exist yet despite being referenced as prepared — it needs to be authored.Scope
docs/the-fleet/index.html— a self-contained, theme-aware "How the Agents Talked" page (how MateMate was built) derived fromdocs/agent-interaction-guide.md, matching the architecture page's design system.docs/index.html— a "How MateMate Works" hub with plain-language, question-based entry points: What is MateMate? → architecture, How was it built? → the-fleet.Constraints
data-theme) design and reduced-motion handling.Done looks like
docs/cross-linked by a shared nav; hub fronts two question-based entry points; both deep pages self-contained and theme-aware; PII scan clean. GitHub Pages publishes on merge to main under/docs.