| /** | |
| * Shopify Market Bridge - Consent-Gated Market Cookie Writer | |
| * | |
| * Theme app extension (app embed) snippet that reads Liquid localization | |
| * data from <script id="shopify-market-data"> and writes it to a cookie | |
| * that the custom pixel sandbox can read via browser.cookie.get(). | |
| * | |
| * Requires storageHelper utility: https://gist.github.com/ceaksan/26c8a59aa8a518732a9a40da967d8cd2 | |
| * | |
| * Consent gate: Shopify.customerPrivacy.analyticsProcessingAllowed() |
| /** | |
| * GTM Custom HTML Tag — Umami purchase event bridge | |
| * File: gtm-umami-purchase-bridge.js | |
| * | |
| * What it does: | |
| * Sends purchase event data + first-party cookie attribution | |
| * to Umami via umami.track(). Runs alongside GA4 purchase tag | |
| * to create an independent, cookie-less attribution record. | |
| * | |
| * Parameters sent to Umami: |
| /** | |
| * GTM Custom JavaScript Variable — page_location + transaction_id | |
| * File: gtm-page-location-transaction-id.js | |
| * | |
| * What it does: | |
| * Appends the ecommerce transaction ID to the current page URL. | |
| * Used as a page_location override in GA4 purchase tags to enable | |
| * order-level debugging in Explore reports. | |
| * | |
| * Example output: |
| /** | |
| * DNM Attribution — Browser console debug script | |
| * File: dnm-debug.js | |
| * | |
| * Usage: | |
| * 1. Open browser console (F12 → Console) | |
| * 2. Paste this script and run it | |
| * 3. Or append ?dnm_debug=1 to any URL (purchase enrichment logs automatically) | |
| * | |
| * Test scenario URLs: |
| /** | |
| * GTM Custom HTML Tag — Purchase source enrichment | |
| * File: gtm-purchase-enrichment.js | |
| * | |
| * What it does: | |
| * When a purchase event fires, reads dnm_src / dnm_cid via storageHelper | |
| * and pushes custom event parameters to dataLayer for the GA4 | |
| * purchase tag to pick up. | |
| * | |
| * Requires: |
| /** | |
| * GTM Custom HTML Tag — First-touch source capture | |
| * File: gtm-source-capture.js | |
| * | |
| * What it does: | |
| * Reads utm_*, gclid, fbclid parameters from the landing URL | |
| * and writes them to two storage keys via storageHelper: | |
| * "dnm_src" and "dnm_cid". | |
| * | |
| * Requires: |
Evaluate every new dependency, pattern, API, or architectural decision against these 8 criteria.
| Criterion | Question | Red Flag |
|---|---|---|
| Benefit | What concrete problem does it solve? Measurable impact? | "Might need it later", "because it's best practice" |
| Necessity | What happens if we don't? Is there a simpler alternative? | Adding something new when existing solution is sufficient |
| { | |
| "$schema": "Instinct JSON Schema — Claude Code Learning Pipeline", | |
| "description": "Schema for mature-instincts.json and global-instincts.json files used by the learning pipeline.", | |
| "raw_capture_record": { | |
| "description": "Written by passive-learner.py (PostToolUse hook) to instincts.jsonl", | |
| "example": { | |
| "ts": "2026-04-07T10:15:00Z", | |
| "project": "my-project", | |
| "pattern": "missing-import", |
| #!/usr/bin/env python3 | |
| """Coffee Analysis CLI — Analyze AI mistake patterns from coffee-log.jsonl. | |
| Usage: | |
| python3 coffee-analyze.py # 7-day summary | |
| python3 coffee-analyze.py --days 30 # 30-day lookback | |
| python3 coffee-analyze.py --format json # JSON output | |
| Data source: ~/.claude/coffee-log.jsonl (produced by coffee-tracker.sh hook) | |
| Debt file: ~/.claude/coffee-debt (cumulative bean count) |