Steel Compendium Workspace

Orchestration repo for the Steel Compendium multi-repo project -- a structured, searchable reference for the Draw Steel TTRPG by MCDM Productions.

Quick Start

git clone --recurse-submodules git@github.com:SteelCompendium/workspace.git steelCompendium
cd steelCompendium
devbox shell
just bootstrap

Authored sub-repos are git submodules; --recurse-submodules fetches them on clone, and just bootstrap is the idempotent catch-up (submodule init + the regenerable data/ scratch dir). For parallel/isolated work use just wt-new <name> — see docs/worktrees-and-submodules.md.

What it looks like

This pipeline produces the live site at steelcompendium.io/v2:

Home — Browse / Read / Bestiary tabs	Bestiary — search/filter + encounter builder

More screenshots (Browse, Read, class pages, statblocks, dark mode) live in v2/README.md.

Layout

steelCompendium/
  justfile              # Workspace recipes (bootstrap, sync, wt-new/wt-rm/wt-status/wt-finish, deploy*)
  devbox.json           # Devbox environment (Go, Node, Python, just, jq, yq, etc.)
  reference/            # Draw Steel condensed reference docs for AI agents
  steel-etl/            # Go ETL pipeline + site builder — THE source of truth for content
                        #   input/{heroes,beastheart,monsters}/*.md = annotated sources
  v2/                   # MkDocs Material site (v2) — built by `steel-etl site`
  compendium/           # MkDocs Material site (v1, deprecated)
  data-gen/             # Legacy ETL pipeline: PDF -> Markdown (deprecated; reference only)
  data-sdk-npm/         # TypeScript SDK for consuming data repos
  draw-steel-elements/  # Web components for Draw Steel content
  statblock-adapter-gl-pages/  # Statblock rendering adapter
  steelCompendium.github.io/   # Root GitHub Pages site (also hosts the SCC API)
  data/                 # Generated output (gitignored; do not edit). bootstrap clones data-unified
    data-unified/       # The single consolidated published data repo:
                        #   en/books/<book>/<format>  (Read: book-faithful, + clean)
                        #   en/unified/<format>        (Browse: cross-book aggregate)

Content flows: annotated steel-etl/input/* → steel-etl gen → data/* → steel-etl site → v2/docs/ → MkDocs build. See ARCHITECTURE.md and steel-etl/README.md.

Recipes

Recipe	Description
just bootstrap	Initialize submodules + data/ scratch (idempotent)
just sync	Pull + move submodules to pinned commits (lockstep)
just wt-new <name>	Create an isolated worktree env (all submodules on branch <name>)
just wt-status <name>	Show an env's submodules ahead + pending pointer bumps
just wt-finish <name>	Land an env's work (pushes submodules + superproject main)
just wt-rm <name>	Tear down an env
just deploy	Full pipeline: gen + SCC API + v2 site
just deploy-api	Pipeline + SCC resolution API only
just deploy-v2	Pipeline + v2 site only