# docs/formal/scraps/

Press-ready LaTeX fragments converted from every document in
`docs/working/`. Each scrap is a self-contained `.tex` fragment (no
`\documentclass`, no `\begin{document}`) ready to `\input{}` into any of the
ten `docs/formal/` publications. See `docs/formal/CLAUDE.md` for the full
scrap header format (`%% SCRAP:`, `%% SOURCE:`, `%% STATUS:`, `%% FITS:`,
`%% EDITORIAL:`) and prose-conversion conventions.

Scraps mirror the source path under `docs/working/`, one subdirectory per
`docs/working/` category:

- **`architecture/`** — from `docs/working/architecture/`.
- **`archive/`** — from `docs/working/archive/`.
- **`experiments/`** — from `docs/working/experiments/`.
- **`hardware/`** — from `docs/working/hardware/`.
- **`papers/`** — from `docs/working/papers/`.
- **`scratch/`** — from `docs/working/scratch/`.
- **`specifications/`** — from `docs/working/specifications/`.
- **`legal/`** — from `docs/working/legal/`. **Legal hold — do not touch.**

A scrap's `FITS:` header lists which publication(s) `\input{}` it; a single
scrap can feed more than one publication (e.g. a physics-engine scrap can
fit both `dev-guide` and `vol1-vm-physics`).

## See also

- [`docs/formal/CLAUDE.md`](../CLAUDE.md) — scrap file naming, header
  format, and editorial-lift conventions.
- [`docs/working/README.md`](../../working/README.md) — the source tier
  these scraps are converted from.
- [Project root](../../../README.md)