Database Explorer

Explore PostgreSQL, MySQL, MongoDB, Redis, and SQLite schemas and sample data safely (read-first). Use when understanding data models or diagnosing query issues.

Navinspire-ia Updated

File contents

Database Explorer

Overview

Map schemas before writing queries. Prefer read-only roles and limits.

Tools

Prefer the built-in db_query tool: it queries SQLite files in the workspace directly (sqlite_path) and named connections (PostgreSQL, Supabase, MySQL, MariaDB) configured under tools.database.connections. Connections are read-only unless allowWrites is set. Fall back to CLI clients (psql, mysql, sqlite3) or MCP servers only when a connection is not configured.

Workflow

  1. Identify engine + connection method (db_query connection, CLI, URL in env, MCP Supabase, etc.).
  2. List databases/schemas/tables - never dump entire prod tables.
  3. Sample with LIMIT / projections; avoid SELECT * on huge tables.
  4. Summarize: entities, keys, relations, row-count estimates.
  5. Flag PII columns; do not paste sensitive rows into chat.

Rules

  • Writes/DDL on shared DBs need human-approval.
  • Credentials from env/secret stores only (secrets-manager).

Navinspire-ia/navin/tree/main/navin/skills/database-explorer commit 5df63c49e9

Frequently asked questions

npx skillmds@latest add navinspire-ia/database-explorer