发现优质的 AI Agent 技能

- ORM is sync-only — `await Model.objects.get()` raises SynchronousOnlyOperation - `sync_to_async` for ORM — `await sync_to_async(Model.objects.get)(pk=1)` - `sync_to_async` default not thread-safe — use `thread_sensitive=True` for ORM - Async view with sync middleware — entire request becomes sync - `database_sync_to_async` in Channels — wrapper for ORM in consumers

- Tests passing in CI — never deploy with failing tests - Environment variables set in target — missing secrets cause silent failures - Database migrations run before code deploy — new code expecting new schema fails - Rollback plan ready — know exactly how to revert before you need to

- Connection pools exhausted = app hangs silently — set max connections, monitor pool usage - Each Lambda/serverless invocation may open new connection — use connection pooling proxy (RDS Proxy, PgBouncer) - Connections left open block schema changes — `ALTER TABLE` waits for all transactions - Idle connections consume memory — set connection timeout, kill idle connections

Use these templates when preparing high-impact database changes.

Use this file on first activation or whenever `~/database-manager/` is missing key files.

Create `~/database-manager/memory.md` with this structure:

Use this file to keep ownership and blast-radius context explicit before major database work.

Use this file when production data integrity or database availability is at risk.

- **Stripe/Paddle:** MRR, churn, LTV, subscriptions via API - **Analytics:** Mixpanel, Amplitude, PostHog, Google Analytics - **CRM:** HubSpot, Pipedrive for sales pipeline - **Custom DB:** Direct PostgreSQL/MySQL queries for user data

group_by: ['alertname', 'service'] ``` **Implement severity hierarchy: P0 (pages immediately) > P1 (within 15min) > P2 (business hours) > P3 (weekly review).** P0: Service completely down, data loss, security breach. P1: Degraded performance, partial outage, high error rates.

Use the Notion API to create/read/update pages, data sources (databases), and blocks.

This skill automates the process of importing employee skills from CSV files into a Supabase database. It parses the CSV, checks what already exists in the database, and generates idempotent SQL scripts to insert missing data.