reversa-docs

Run the Reversa core extraction first to produce soul.md, chronicle.md, and SDD specs, then invoke /reversa-docs. It orchestrates four agents that render architecture, metrics, glossary, and feature pages into _reversa_docs/ as a self-contained mini-site.

It produces a mini-site in _reversa_docs/ containing arquitetura.html, modulos.html, metricas.html, timeline.html, glossario.html, a deck, per-feature pages, and an index.html with navigation. Pages are omitted gracefully when their source artifacts are missing.

Yes. The Publisher downloads Three.js, D3, and Highcharts into assets/vendor/ and embeds all data in assets/js/data.js, so index.html works via double-click over file:// with no fetch calls or CDN references.

Yes. On a second run you can regenerate everything, a single agent's pages, or one specific page. Each agent is also invocable standalone via /reversa-docs-<name>, and automatic backups are created in .backup-<timestamp>/ before any overwrite.

The skill detects missing sources and offers to either run /reversa first or generate a minimal index page with only the seal. Pages depending on absent sources are omitted with recorded reasons rather than failing.

It never writes outside _reversa_docs/ and only reads core artifacts, so it cannot extract knowledge itself. It depends on prior Reversa extraction output, and visual reproducibility relies on a sha256 seed derived from soul.md or the project name.