diff --git a/CATALOG.md b/CATALOG.md new file mode 100644 index 0000000..a0a52e2 --- /dev/null +++ b/CATALOG.md @@ -0,0 +1,41 @@ +--- +title: "Agentic Semantic Web" +description: "CSS framework for agent-authored web content — semantic HTML, data-attributes, zero classes" +status: release +state: published +created: 2026-04-24 +tags: + - release + - framework + - css +repo: trentuna/asw +authors: + - name: Ludo + role: Author + - name: Vigo + role: Maintainer +--- + +## Context + +The Agentic Semantic Web (ASW) is a CSS and HTML framework purpose-built for agent-authored web content. It treats markup as a communication medium between autonomous systems and human readers — every element is semantic, every data attribute carries meaning, and the styling is a layer on top, not the source of truth. + +ASW is a three-layer stack: + +1. **Pico CSS** — minimal residual styles as the base layer. A lightweight, class-light semantic HTML framework that provides sensible defaults for headings, forms, tables, and typography without requiring any custom classes. +2. **Theme tokens** — CSS custom properties that define the visual identity. Colors, spacing, typography scale, and interaction states are all exposed as `--asw-*` variables. +3. **Data-attribute extensions** — Custom properties and attribute selectors that extend the base framework with agent-oriented patterns: `[data-asw-type]`, `[data-asw-status]`, `[data-asw-agent]` for annotating machine-generated content. + +Zero classes. No `.btn`, no `.card`, no `.container`. Everything is semantic HTML elements styled through extended data-attribute selectors. The framework assumes the author is a language model that can write semantic HTML but shouldn't have to remember a framework's arbitrary class naming convention. + + + +## Related + +- [OpenCD](./../projects/opencd/CATALOG.md) — built on ASW philosophy +- [Human Advisory](./../projects/human-advisory/CATALOG.md) — ships with ASW as embedded component + +## Credits + +- **Ludo** — Author, concept +- **Vigo** — Maintainer, distribution