Add AD1 session data, memory entries for datasheet pipeline and security incident

- Imported AD1 Claude session files to clients/dataforth/session-logs/
- Created memory: project_datasheet_pipeline.md (full pipeline architecture)
- Created memory: project_dataforth_incident_2026-03-27.md (security incident + MFA)
- Updated MEMORY.md index
- Updated session log with AD1 pipeline rebuild findings

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

clients/dataforth/session-logs/project_pipeline_rebuilt.md

@@ -0,0 +1,47 @@
+---
+name: Test datasheet pipeline rebuilt 2026-03-27
+description: New server-side pipeline replaces CTONWTXT/DFWDS/TestDataSheetUploader. Generates exact-match TXT from DAT imports. 1470 model specs loaded.
+type: project
+---
+
+## What Was Built (2026-03-27)
+
+All files in C:\Shares\testdatadb\:
+
+### New Files
+- `parsers/spec-reader.js` — Reads 8 binary spec DAT files, 1470 models across 6 families (SCM5B, 8B, DSCA, DSCT, SCM7B, plus sub-families 5B45, 5B48)
+- `templates/datasheet-exact.js` — Exact-match TXT formatter reverse-engineered from QuickBASIC DATASHEETWRITE
+- `database/export-datasheets.js` — Standalone/library export script, writes {SN}.TXT to X:\For_Web
+- `database/archive-for-web.js` — Archives old For_Web files into year subfolders
+- `database/generate-customer-pdfs.js` — One-off script for Quatronix customer order (72/73 generated)
+- `specdata/` — Binary spec DAT files (5BMAIN, 8BMAIN, DSCOUT, SCTMAIN, DSCMAIN4, 7BMAIN, 5B45DATA, DB5B48)
+
+### Modified Files
+- `database/import.js` — Post-import hook calls export-datasheets after incremental imports
+- `database/schema.sql` — Added datasheet_exported_at and forweb_exported_at columns
+- `routes/api.js` — Added /api/datasheet/:id/pdf endpoint (pdfkit)
+- `public/index.html` — Added ALL filter, auto-search on load, PDF download button, export status in detail view
+- `package.json` — Added pdfkit dependency
+
+### Database Changes
+- `datasheet_exported_at` — When TXT was generated
+- `forweb_exported_at` — When file was confirmed in X:\For_Web
+- 1,436,007 records pre-marked as exported (matched existing For_Web files)
+
+### Pipeline Flow
+DAT file arrives via NAS sync → import.js parses to DB → export-datasheets generates {SN}.TXT → X:\For_Web
+
+### What's Eliminated
+- CTONWTXT.BAT (DOS-side TXT transfer) — no longer needed
+- DFWDS.exe (filename decoder/validator) — no longer needed
+- TestDataSheetUploader (VB.NET web uploader) — endpoints are 404, needs replacement
+
+### Remaining Gaps
+- Website upload: old ASP.NET endpoints dead (404), need new upload mechanism
+- SCM5B49 spec file empty (0 bytes) — need John's help
+- 7B datasheet formatting not yet added to datasheet-exact.js (7B has 30+ unique parameters)
+- Service runs as SYSTEM causing permission issues with SHM/WAL files
+- 500K files archived into year subfolders in X:\For_Web
+
+**Why:** Rebuilt to recover from crypto wipe and unblock customer shipments (Quatronix/China blocking on missing datasheets).
+**How to apply:** For new test data, the pipeline is automatic via Sync-FromNAS. For the 830K backlog of unexported records, run `node database/export-datasheets.js --limit N`.