postgres-intelligencelisted

# postgres-intelligence PostgreSQL intelligence for LLM coding agents. It gives Claude Code, Codex, and other local agents a credential-safe way to discover PostgreSQL schemas, run read-first SQL, inspect query errors, and reason about performance. The core rule is simple: the agent should not open or print `.env` directly. Scripts load credentials at runtime, and the agent only sees safe summaries, query results, and metadata. ## What It Does - Loads one or more PostgreSQL connections from `.env` using `DB1_...DB10_...`. - Tests connectivity without printing passwords or full DSNs. - Extracts schema metadata from `information_schema` and `pg_catalog`. - Executes read-only SQL by default: `SELECT`, `WITH`, `SHOW`, and `EXPLAIN`. - Blocks writes and DDL unless explicit flags are passed after user approval. - Returns structured JSON that any LLM agent can parse. - Provides PostgreSQL-specific guidance for indexes, JSONB, `EXPLAIN`, and maintenance. ## Install ```bash cd postgres-intelligence python3 -m venv .venv . .venv/bin/activate python -m pip install -r requirements.txt cp .env.example .env python scripts/config.py python scripts/db_connector.py python scripts/schema_extractor.py ``` Use `.env.example` as the public template. Keep real `.env`, `.venv`, and `schema_metadata.json` out of git. ## Configure ```bash DB1_HOST=localhost DB1_PORT=5432 DB1_USER=your_username DB1_PASSWORD=your_password DB1_DATABASE=your_database DB1_NAME=primary DB1_SSLMODE=prefer DB1_CONNE