Seed JSON files (parsed legislation) are now committed so that
npm run build:db works on any machine without re-ingestion.
Raw source HTML and .db files remain gitignored.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Complete production implementation with shell+adapter architecture,
13 MCP tools, SQLite FTS5 search, and multi-source ingestion pipeline.
Ingestion fetches from UCI mirror, UNODC SHERLOC PDFs, and Knesset
mobile PDFs (135 provisions, 33 definitions). 3 acts with full text,
7 acts metadata-only due to gov.il/nevo.co.il access restrictions.
Knesset OData API used for metadata enrichment.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
