feat: Tier B operator scaffolding — bundle, copy SoT, posts, emails

Pick up and finish yesterday's cut-off Tier B pass.

- build/: PyInstaller scaffold (datatools.spec + launcher.py +
  hook-streamlit.py + README) — folder-mode bundle, locked
  127.0.0.1, per-OS recipe
- marketing/COPY.md: single source of truth for every customer-facing
  string — landing H1/sub/CTAs, demo CTAs, email subjects, Gumroad
  listing, banned phrases
- marketing/community-posts/: 9 drafts (3 posts × 3 niches:
  bookkeeper, revops, shopify-pet) — story / tip / soft-offer
- marketing/emails/: 18 drafts (Gumroad delivery + 5-touch
  onboarding × 3 niches), per-niche segmentation guidance
- docs/NEXT-STEPS.md: flip 2.2 / 2.4 / 3.1 / 3.4 to done with
  pointers to the new assets; add Phase 0 inventory rows
- .gitignore: narrow `build/` ignore so PyInstaller spec + launcher
  + hooks get tracked, only generated artifacts (build/build/,
  build/__pycache__/, build/dist/) stay ignored

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

marketing/emails/bookkeeper/04-day14.md

@@ -0,0 +1,35 @@
+# Bookkeeper · Day 14 — Two-minute trick: the gate report
+
+**Subject:** Two-minute trick: the gate report
+**Send:** Day 14
+**Goal:** surface the gate tool — non-obvious, high-value once seen
+
+---
+
+Hi {{first_name}},
+
+The tool inside DataTools that buyers find last is the **gate** — and it's the one that quietly does the most for you.
+
+What it does: before any row gets written to the cleaned CSV, the gate runs a per-row pass-through check. Rows that fail get *quarantined* into a separate file (`<filename>.quarantine.csv`) instead of silently dropped or silently passed.
+
+Default rules (you can add your own):
+
+- Missing required fields (date, amount)
+- Amount in unexpected currency without a flag
+- Date outside the export's stated range (catches the "totals row" issue from Day 1)
+- Duplicate of another row already in the file (per the dedupe pass)
+- Confidence below your threshold on a field that got auto-corrected
+
+The 2-minute workflow:
+
+1. Run the pipeline as usual.
+2. Open `<filename>.quarantine.csv`. (It'll be tiny — typically 0-5% of rows.)
+3. Eyeball it. Anything that's a real transaction, fix-and-re-include manually. Anything that's a totals row / blank row / corrupt row — confirm it's correctly quarantined and delete it.
+4. Re-run the pipeline on the fixed-up version (or just append the manually-fixed rows to the cleaned CSV).
+
+The reason this matters: silent drops are the worst possible failure mode for a bookkeeper. You'd rather a row come out wrong (you'll catch it on review) than disappear (you won't catch it for months). The gate makes the silent-drop case impossible.
+
+Set the gate's confidence threshold to `0.85` for client work. Lower (0.75) for personal / exploratory; higher (0.92+) only if you've spent time tuning your client's preset.
+
+— Michael
+{{support_email}}