Ingestion Pipeline

Pipeline sequence for each new source document:

1. New source file is placed in ../current-documents or ../input-documents.
2. Manifest metadata is updated in ingestion/manifests/current-mvp.json (or a new snapshot manifest).
3. npm run import:documents extracts markdown from the canonical source format.
4. Normalization applies clause anchors and removes register export artifacts.
5. Full text and section pages are generated in src/content/docs/....
6. Original files and markdown exports are mirrored to public/downloads/....
7. Link suggestion files are generated for human review before publication.
8. Human review confirms legal interpretation and cross-links.

Human review is mandatory

The import pipeline is an acceleration layer. It does not replace legal review.