dataflow

Backend Factory - Data Entity Lifecycle

Trace a single data entity through every layer of the backend to map its complete lifecycle — from birth to death.

Step 1: Parse Arguments

$ARGUMENTS should contain a data entity name (e.g., User, Order, Payment, Session).

If $ARGUMENTS is empty, scan the codebase for model/schema definitions and list discovered entities.

Step 2: Detect Framework & ORM

Quick check for web framework, ORM/DB layer, cache layer, and queue system.

Step 3: Search for ALL References

Cast a wide net across all layers:

• DB Schema / Model Definitions: model files, migrations, schema definitions
• Route Handlers (CRUD): create, read, update, delete endpoints
• Services & Business Logic: transformation, validation, state machines
• Cache References: cache keys, invalidation, TTLs
• Queue Jobs: background processing, bulk operations
• Webhook Handlers: incoming webhooks that create/update the entity
• Relationships: foreign keys, join tables, embedded documents

Step 4: Map Complete Lifecycle

Organize references into lifecycle stages:

• Birth: creation triggers, validations, side effects
• Storage: primary/secondary stores, file storage
• Transformation: mutation paths, state transitions, audit trails
• Access: read paths, authorization checks
• Relationships: parent, child, peer entities, cascading operations
• Death: soft delete, hard delete, GDPR erasure, TTL expiry, archival

Step 5: Generate Visualization

If the factory server is running, POST lifecycle overlay data.

Step 6: Report with Narration

Present findings in two formats:

Technical Summary

Entity details, lifecycle stages with file paths, relationship map, potential issues.

First-Person Narration (from the entity's perspective)

A creative narration told BY the data entity as a factory worker. Use real file names, routes, and field names.

Suggest: "Run /backend-factory:trace on specific routes to see the detailed station-by-station path."