PostgreSQL best-practices skill for AI coding agents: schema design, indexing, query optimization, migrations, locking, security and Row-Level Security, connection pooling, MVCC/VACUUM, WAL, replication, monitoring, and PlanetScale-specific operations.

One skill, 32 reference files. The agent loads SKILL.md (workflow, core rules, recommendation format) and pulls in only the reference it needs.

npx skills add PyModel/postgres-skill

Or symlink skills/postgres into your agent's skills directory (~/.claude/skills, ~/.codex/skills, ~/.pi/agent/skills, ~/.agents/skills, ...).

skills/postgres/
├── SKILL.md          # workflow, core rules, recommendation format, reference index
└── references/       # one topic per file, loaded on demand

Forked from planetscale/database-skills (skills/postgres) and merged with an evidence-driven best-practices skill: locking, security/RLS, migrations, EXPLAIN, JSONB/arrays, full-text search, window functions/CTEs, and CTID batch processing were added; overlapping and outdated guidance was reconciled and every factual claim was re-verified against the PostgreSQL 18 documentation.

MIT licensed — see LICENSE.