Dynamic, Schemaless Work Item Tracking for SunShade
Eliminating rigid database ENUMs in favor of dynamic per-project JSON schemas. Programmatic headless ingestion, polymorphic work items, and fractional indexing for modern collaborative tools.
Schemaless Polymorphism
No rigid PostgreSQL ENUMs. Work item types, statuses, hierarchy levels, and metadata are validated against dynamic JSON schemas stored in project settings.
Gemini Spark Ingest API
Headless POST /api/v1/items/ingest endpoint. External LLMs can programmatically create, nest, and upsert work items without direct DB access.
Fractional Reordering
Midpoint double precision ordering allows cards to be dropped and reordered anywhere in Kanban columns without cascading table updates or locks.
External pipelines authenticate with their tenant API key and pass batch payloads. Parents are resolved dynamically via parent_ref_id:
// Headers:
// Authorization: Bearer tk_live_sunshade_master_key
// Content-Type: application/json
POST https://track.sunshade.icu/api/v1/items/ingest
{
"project_slug": "portfolio",
"items": [
{
"external_ref_id": "SPEC-HUB-11",
"title": "Deploy Sovereign Event Bus to PatchWork",
"description": "Auto-generate maintenance tasks from citizen reports.",
"item_type": "story",
"status": "in_progress",
"assignee": "tympollack",
"metadata": {
"complexity": 3,
"priority": "High",
"origin_agent": "Gemini Spark"
}
},
{
"external_ref_id": "TASK-HUB-11-A",
"parent_ref_id": "SPEC-HUB-11",
"title": "Implement POST /api/events Webhook Route",
"item_type": "task",
"status": "planned",
"metadata": {
"complexity": 1
}
}
]
}