Can I vibe code Feedops?

Build a standalone Product Feed Generator microservice for Google Shopping and Meta Ads using Node.js, Express, and PostgreSQL.

1. DATA MODEL:
- Stores: store_id, platform (shopify/custom), api_credentials, sync_frequency.
- Products: product_id, store_id, raw_payload (JSONB), standardized_payload (JSONB), sync_status, updated_at.
- FeedConfigs: feed_id, store_id, channel (google/meta), title_template, field_mappings (JSONB), filter_rules (JSONB).

2. CATALOG SYNC & TRANSFORM ENGINE:
- Build an ingestion job that fetches catalog data via Shopify GraphQL API using pagination cursor loops.
- Store raw inventory, prices, images, variants, and product types in PostgreSQL.
- Build a transformation worker that parses raw product data and applies rules:
  * Title template parsing (e.g., '{brand} {title} {color} {size}').
  * Standardize variant mapping: set g:item_group_id to parent product ID, assign g:availability based on inventory level.
  * Map shop categories to Google Product Taxonomy strings via a static lookup dictionary.

3. FEED GENERATION & OUTPUT:
- Implement streaming XML generation using 'fast-xml-parser' or stream writers to handle up to 50,000 SKUs without memory exhaustion.
- Format output to adhere strictly to Google Merchant Center XML specification (rss 2.0 with g: namespace).
- Save generated XML to a public S3 bucket or serve via a cached express GET route with proper headers (application/xml).

4. OPTIONAL LLM ENRICHMENT:
- Add a background task calling OpenAI GPT-4o-mini to rewrite product titles missing key attributes like material or gender before feed serialization.

5. FAILURE MODES & MONITORING:
- If product missing image_link or price, flag record with validation_error and skip feed inclusion.
- Return non-200 standard log if feed generation exceeds execution threshold.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs