A Node.js script that scans the repository's Markdown files (reports, posts, contributors), extracts structured metadata (e.g., sections, recommenders, reading time), and builds a full-text search index using MiniSearch. It demonstrates a reusable pattern for turning a collection of markdown documents into a searchable archive, handling CJK tokenization and diverse content models. Although the file paths are specific to this project, the architecture (parse -> extract -> index) is easily adaptable to other static content sites.

A GitHub Actions workflow that automates the build and deployment of the static site to GitHub Pages. It installs Node.js, runs `npm run build` (which invokes `build-index.js`), and then deploys the `docs/` directory to Pages. It defaults to a self-hosted runner with GNU tar. This is a standard but clean example of a CI pipeline for a static site generator, and the pattern (install dependencies, build, deploy) is reusable for any similar project.

A single-file SPA that provides a hash-based router, markdown rendering (via Marked), and full-text search (via MiniSearch). It lazy-loads content from `index.json` and `search.json` generated by the build script. The application is self-contained, responsive, and includes features like dark mode and keyboard shortcuts. While the content loading is specific to the project's data format, the SPA architecture and search integration can serve as a reference for building a static site with client-side search.