agentsclimarketplace

Postgres patterns

Skill the-hugin/RSIm/skills/postgres-patterns

Recursively Self-Improving Module — persistent memory + structured improvement loop for Claude Code

Install
npx -y skills add the-hugin/RSIm --skill postgres-patterns

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

PostgreSQL best practices for query optimization, schema design, indexing, RLS, and connection management. Includes Drizzle ORM patterns for TypeScript — schema definition, type inference, query builder style. Based on Supabase and lobehub guidelines.

SKILL.md

10.2 KB, ~2.6k tokens by cl100k_base, as published. Nobody here has run it

PostgreSQL Patterns

Quick reference for PostgreSQL best practices.

When to Activate

  • Writing SQL queries or migrations
  • Designing database schemas
  • Troubleshooting slow queries
  • Implementing Row Level Security
  • Setting up connection pooling
  • Defining Drizzle ORM schemas and relations (TypeScript)

Index Cheat Sheet

Query PatternIndex TypeExample
WHERE col = valueB-tree (default)CREATE INDEX idx ON t (col)
WHERE col > valueB-treeCREATE INDEX idx ON t (col)
WHERE a = x AND b > yCompositeCREATE INDEX idx ON t (a, b)
WHERE jsonb @> '{}'GINCREATE INDEX idx ON t USING gin (col)
WHERE tsv @@ queryGINCREATE INDEX idx ON t USING gin (col)
Time-series rangesBRINCREATE INDEX idx ON t USING brin (col)

Data Type Quick Reference

Use CaseCorrect TypeAvoid
IDsbigintint, random UUID
Stringstextvarchar(255)
Timestampstimestamptztimestamp
Moneynumeric(10,2)float
Flagsbooleanvarchar, int

Common Patterns

Composite Index Order

-- Equality columns first, then range columns
CREATE INDEX idx ON orders (status, created_at);
-- Works for: WHERE status = 'pending' AND created_at > '2024-01-01'

Covering Index (avoids table lookup)

CREATE INDEX idx ON users (email) INCLUDE (name, created_at);
-- SELECT email, name, created_at — no heap fetch needed

Partial Index (smaller, faster)

CREATE INDEX idx ON users (email) WHERE deleted_at IS NULL;

RLS Policy (optimized)

CREATE POLICY policy ON orders
  USING ((SELECT auth.uid()) = user_id);  -- Wrap in SELECT!

UPSERT

INSERT INTO settings (user_id, key, value)
VALUES (123, 'theme', 'dark')
ON CONFLICT (user_id, key)
DO UPDATE SET value = EXCLUDED.value;

Cursor Pagination (O(1) vs OFFSET O(n))

SELECT * FROM products WHERE id > $last_id ORDER BY id LIMIT 20;

Queue Processing (skip locked)

UPDATE jobs SET status = 'processing'
WHERE id = (
  SELECT id FROM jobs WHERE status = 'pending'
  ORDER BY created_at LIMIT 1
  FOR UPDATE SKIP LOCKED
) RETURNING *;

Diagnostics

-- Find unindexed foreign keys
SELECT conrelid::regclass, a.attname
FROM pg_constraint c
JOIN pg_attribute a ON a.attrelid = c.conrelid AND a.attnum = ANY(c.conkey)
WHERE c.contype = 'f'
  AND NOT EXISTS (
    SELECT 1 FROM pg_index i
    WHERE i.indrelid = c.conrelid AND a.attnum = ANY(i.indkey)
  );

-- Find slow queries (requires pg_stat_statements)
SELECT query, mean_exec_time, calls
FROM pg_stat_statements
WHERE mean_exec_time > 100
ORDER BY mean_exec_time DESC;

-- Check table bloat
SELECT relname, n_dead_tup, last_vacuum
FROM pg_stat_user_tables
WHERE n_dead_tup > 1000
ORDER BY n_dead_tup DESC;

Configuration Template

-- Connection limits (tune for RAM)
ALTER SYSTEM SET max_connections = 100;
ALTER SYSTEM SET work_mem = '8MB';

-- Timeouts
ALTER SYSTEM SET idle_in_transaction_session_timeout = '30s';
ALTER SYSTEM SET statement_timeout = '30s';

-- Monitoring
CREATE EXTENSION IF NOT EXISTS pg_stat_statements;

-- Security defaults
REVOKE ALL ON SCHEMA public FROM public;

SELECT pg_reload_conf();

Anti-Patterns

Anti-PatternProblemFix
OFFSET N paginationO(n) table scanCursor pagination with id > $last
varchar(255)Arbitrary limit, no benefitUse text
float for moneyRounding errorsnumeric(10,2)
timestampNo timezone infotimestamptz
Missing FK indexesSlow JOINsIndex all foreign keys
SELECT * in RLSForces expression eval per rowWrap auth calls in SELECT subquery

Based on Supabase Agent Skills — MIT License


Drizzle ORM (TypeScript)

Применяй когда схема и запросы пишутся на TypeScript через Drizzle ORM.

Config

// drizzle.config.ts
export default defineConfig({
  dialect: 'postgresql',
  schema: './src/database/schemas/*',
  out: './src/database/migrations',
  strict: true,
});

Naming Conventions

ОбъектФорматПример
Таблицыplural snake_caseusers, session_groups
Колонкиsnake_caseuser_id, created_at

Column Patterns

Primary Key — prefixed text ID (читаемый, distinguishable по типу):

id: text('id')
  .primaryKey()
  .$defaultFn(() => idGenerator('agents'))
  .notNull(),
// Для внутренних таблиц: uuid вместо text

Foreign Key с cascade:

userId: text('user_id')
  .references(() => users.id, { onDelete: 'cascade' })
  .notNull(),

Timestamps — через хелперы, не вручную:

// _helpers.ts
export const createdAt = () => timestamptz('created_at').defaultNow().notNull();
export const updatedAt = () => timestamptz('updated_at').$onUpdate(() => new Date());
export const timestamps = { createdAt: createdAt(), updatedAt: updatedAt() };

// В схеме:
...timestamps,

Indexes — возвращай массив (не объект):

(t) => [uniqueIndex('client_id_user_id_unique').on(t.clientId, t.userId)],

Full Table Example

export const agents = pgTable(
  'agents',
  {
    id: text('id')
      .primaryKey()
      .$defaultFn(() => idGenerator('agents'))
      .notNull(),
    slug: varchar('slug', { length: 100 })
      .$defaultFn(() => randomSlug(4))
      .unique(),
    userId: text('user_id')
      .references(() => users.id, { onDelete: 'cascade' })
      .notNull(),
    clientId: text('client_id'),
    config: jsonb('config').$type<AgentConfig>(),
    ...timestamps,
  },
  (t) => [uniqueIndex('client_id_user_id_unique').on(t.clientId, t.userId)],
);

Type Inference

export const insertAgentSchema = createInsertSchema(agents);
export type NewAgent = typeof agents.$inferInsert;
export type AgentItem  = typeof agents.$inferSelect;

Junction Table (Many-to-Many)

export const agentsKnowledgeBases = pgTable(
  'agents_knowledge_bases',
  {
    agentId: text('agent_id')
      .references(() => agents.id, { onDelete: 'cascade' }).notNull(),
    knowledgeBaseId: text('knowledge_base_id')
      .references(() => knowledgeBases.id, { onDelete: 'cascade' }).notNull(),
    userId: text('user_id')
      .references(() => users.id, { onDelete: 'cascade' }).notNull(),
    enabled: boolean('enabled').default(true),
    ...timestamps,
  },
  (t) => [primaryKey({ columns: [t.agentId, t.knowledgeBaseId] })],
);

Query Style — критическое правило

Всегда db.select(). Никогда db.query.*.

db.query.findMany/findFirst/with: генерирует сложные lateral joins с json_build_array — хрупкие и трудно отлаживаемые.

Select single row:

// ✅
const [result] = await db.select().from(agents).where(eq(agents.id, id)).limit(1);

// ❌
return db.query.agents.findFirst({ where: eq(agents.id, id) });

Select with JOIN:

// ✅
const rows = await db
  .select({
    runId: evalRunTopics.runId,
    score: evalRunTopics.score,
    testCase: evalTestCases,
  })
  .from(evalRunTopics)
  .leftJoin(evalTestCases, eq(evalRunTopics.testCaseId, evalTestCases.id))
  .where(eq(evalRunTopics.runId, runId))
  .orderBy(asc(evalRunTopics.createdAt));

// ❌
return db.query.evalRunTopics.findMany({
  where: eq(evalRunTopics.runId, runId),
  with: { testCase: true },
});

Aggregation:

// ✅
const rows = await db
  .select({
    id: datasets.id,
    name: datasets.name,
    count: count(testCases.id).as('count'),
  })
  .from(datasets)
  .leftJoin(testCases, eq(datasets.id, testCases.datasetId))
  .groupBy(datasets.id);

One-to-Many — два отдельных запроса:

// ✅ Два простых запроса вместо relational with:
const [parent] = await db.select().from(datasets).where(eq(datasets.id, id)).limit(1);
if (!parent) return undefined;

const children = await db
  .select()
  .from(testCases)
  .where(eq(testCases.datasetId, id))
  .orderBy(asc(testCases.sortOrder));

return { ...parent, testCases: children };

UPSERT:

await db
  .insert(settings)
  .values({ userId, key, value })
  .onConflictDoUpdate({
    target: [settings.userId, settings.key],
    set: { value: sql`excluded.value` },
  });

Cursor pagination в Drizzle:

const rows = await db
  .select()
  .from(products)
  .where(gt(products.id, lastId))
  .orderBy(asc(products.id))
  .limit(20);

Транзакции:

await db.transaction(async (tx) => {
  const [user] = await tx.insert(users).values(userData).returning();
  await tx.insert(profiles).values({ userId: user.id, ...profileData });
});

Drizzle Anti-Patterns

Anti-PatternПроблемаFix
db.query.* с with:Lateral joins, хрупкоdb.select() + leftJoin()
One-to-many через with:Сложный JSON в памятиДва отдельных запроса
Indexes как объектDeprecated синтаксисВозвращай массив (t) => [...]
Ручные timestamp поляНесогласованностьХелперы createdAt(), updatedAt()
$inferInsert не экспортируетсяДублирование типов вручнуюВсегда экспортируй NewX и XItem

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 3 of the 12 instructions most databases sql skills give in ~2.6k tokens

Counted across 589 of the 662 authors here whose files we hold, read 2026-08-07

  • Use parameterized queriesin 37 of 589, across 34 files
  • Use timestamptz for timestampshere, and in 30 of 589, across 14 files
  • Index foreign keyshere, and in 29 of 589, across 18 files
  • Create indexes concurrentlyin 29 of 589, across 24 files
  • Use numeric type for moneyin 25 of 589, across 8 files
  • Use cursor pagination instead of offsethere, and in 24 of 589, across 17 files
  • Select only required columnsin 24 of 589, across 20 files
  • Add indexes manually on foreign key columnsin 22 of 589, across 12 files
  • Normalize to third normal formin 19 of 589, across 10 files
  • Configure connection poolingin 19 of 589, across 17 files
  • Put equality columns before range columns in indexesin 18 of 589, across 10 files
  • Read individual rule files for detailed explanationsin 18 of 589, across 4 files

Said here and by no other author read

  • define indexes as an array
  • use timestamp helper functions
  • export inferred insert and select types
  • use db.select instead of db.query
  • fetch one-to-many relations using two separate queries

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.