Agent-Skills.md

Agent Skills: Supabase & Backend Architecture Skill

Expert in Supabase architecture, SQL optimization (PostgreSQL), and backend security (RLS) for real-time tracking systems.

UncategorizedID: abelv22/project-foundation/supabase-backend

Install this agent skill to your local

pnpm dlx add-skill https://github.com/AbelV22/project-foundation/tree/HEAD/.agent/skills/supabase-backend

Skill Files

Browse the full folder contents for supabase-backend.

Download Skill

Loading file tree…

.agent/skills/supabase-backend/SKILL.md

Skill Metadata

Name
supabase-backend
Description
Expert in Supabase architecture, SQL optimization (PostgreSQL), and backend security (RLS) for real-time tracking systems.

Supabase & Backend Architecture Skill

This skill enables the assistant to provide high-level architectural advice and implementation details for the iTaxiBcn backend.

Knowledge Areas

1. Database Schema Optimization

  • Time-Series Data: Guidelines for handling high-frequency location updates in registros_reten and geofence_logs.
  • Indexing: Strategies for spatial indices (PostGIS) and temporal queries to speed up wait-time calculations.
  • Materialized Views: Recommendation for replacing heavy queries on registros_reten with materialized views for zone aggregations.

2. Row Level Security (RLS)

  • Device-Based Access: Ensuring device_id based security since the app currently uses device identifiers instead of full user auth (until Phase 2).
  • Audit Logs: Best practices for geofence_logs and location_debug_logs security.

3. Edge Functions (Deno/TypeScript)

  • Geofencing Logic: Optimization of the Ray-casting algorithm in check-geofence.
  • Performance: Minimizing startup time and memory footprint of edge functions.
  • Error Handling: Robust try-catch patterns and standard JSON responses.

4. SQL Scripting

  • Migrations: Following the supabase/migrations/ structure.
  • Stored Procedures: Writing efficient PL/pgSQL for complex logic like useWhereNext score calculation on the server side.

Guidelines for Responses

  • Always suggest Materialized Views for dashboard metrics that don't need second-by-second accuracy.
  • When writing SQL, ensure idempotency (use CREATE OR REPLACE or IF NOT EXISTS).
  • Prioritize PostGIS functions for distance and polygon math.