菜单
Server PostgreSQL operations — connect, migrate, Timescale, pgvector, introspection; delegates deep tuning to postgres-optimization.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/SimHacker/moollm --skill postgres命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | postgres |
| description | Server PostgreSQL operations — connect, migrate, Timescale, pgvector, introspection; delegates deep tuning to postgres-optimization. |
| license | MIT |
| tier | 1 |
| allowed-tools | ["read","grep","glob","run_terminal_cmd"] |
| related | ["postgres-optimization","sqlite","schema","robust-first","plan-then-execute"] |
| benefits_from | ["postgres-optimization"] |
| tags | ["moollm","postgres","sql","timescale","pgvector","migrations","psql"] |
Part of MOOLLM · skills/postgres/
Treat this skill as the MOOLLM wrapper around PostgreSQL’s full stack: the psql client, pg_dump / pg_restore, connection libraries, extension packages, and the on-disk / managed-server ecosystem—not a reimplementation of the engine. The postgres schemapedia mechanism points here; sql remains the abstract language family.
This skill covers operating PostgreSQL: getting connected, migrations, extensions (including TimescaleDB and pgvector), catalog introspection, and operational vocabulary. Index design, constraint tricks, and deep EXPLAIN work live in postgres-optimization — use this skill first for wiring and safety, then delegate when the problem is plan-level.
postgres-optimization after you have a query and schema.Single URI (typical in apps):
postgresql://user:password@host:5432/dbname?sslmode=require
psql accepts a URI:
psql "$DATABASE_URL" -v ON_ERROR_STOP=1 -c "
SELECT current_database(), current_user, version();
"
If tooling uses PG_CONNECTION_STRING instead of DATABASE_URL, normalize in shell:
psql "${DATABASE_URL:-$PG_CONNECTION_STRING}" -v ON_ERROR_STOP=1 -f ./migrations/001_apply.sql
Split variables (classic DBA style): PGHOST, PGPORT, PGUSER, PGDATABASE, PGPASSWORD. psql picks these up when no URI is passed.
-v ON_ERROR_STOP=1 — stop the script on first error (essential for migrations).CREATE TABLE IF NOT EXISTS, CREATE EXTENSION IF NOT EXISTS — avoid blind rewrites in production without review.Version and extensions:
SELECT version();
SELECT extname, extversion FROM pg_extension ORDER BY 1;
Table sizes (rough):
SELECT relname, pg_total_relation_size(oid) AS bytes
FROM pg_class
WHERE relkind = 'r' AND relnamespace = 'public'::regnamespace
ORDER BY bytes DESC
LIMIT 20;
Load pg_stat_statements (if installed) and read the PostgreSQL docs for setup — then inspect top queries by total time.
Timescale extends PostgreSQL with hypertables (time-partitioned tables), compression, retention policies, and continuous aggregates (rollup views). Read the product docs before turning this on in production: chunk intervals, retention, and compression interact with query patterns.
Canonical entry points:
Typical enable: CREATE EXTENSION IF NOT EXISTS timescaledb; then create_hypertable(...) per current API in your Timescale version.
pgvector stores vectors in-table and supports nearest-neighbor search with distance operators. Install the extension package for your Postgres major, then:
CREATE EXTENSION IF NOT EXISTS vector;
Details and operator list: https://github.com/pgvector/pgvector
See EXTENSIONS.md for a link table (Timescale, pgvector, pg_stat_statements, pg_trgm, PostGIS, and core manual links).
pg_dump / pg_restore — portable, version-sensitive.| Skill | Role |
|---|---|
| postgres-optimization | Unconventional and advanced performance (indexes, EXPLAIN, plans). |
| sqlite | Embedded file database — when to stay local vs move to Postgres. |
| schema | Schemapedia registry — sql language family, postgres mechanism. |
postgres-optimization — use when this skill’s connection and ops context is not enough and the bottleneck is query or index design.sqlite — compare embedded vs server tradeoffs.schema — mechanism postgres in the registry points here.This skill's directory (browse and fetch everything): skills/postgres/