Skill Detail

Expose guarded database connections to agents with Data Peek

Use Data Peek's local MCP server to let agents inspect SQL databases through capped read tools and human-approved writes.

Data Extraction & TransformationMCP
Data Extraction & Transformation MCP Security Reviewed
⭐ 1.7k GitHub stars ⬇ 99/wk npm
COPY SKILL INSTRUCTIONS (OPTIONAL)
npx skills add agentskillexchange/skills --skill expose-guarded-database-connections-to-agents-with-data-peek Copy
Uses the third-party skills CLI, not an ASE-owned installer. Check your agent’s compatibility. This copies instructions; complete the upstream tool setup below separately.
At a glance
Tools required
Data Peek desktop app, built-in MCP server, MCP-capable agent client, and a configured PostgreSQL, MySQL, SQL Server, SQLite, or ClickHouse connection.
Install & setup
Download Data Peek from https://www.datapeek.dev/ or build from source with pnpm install and pnpm dev. Configure a database connection in the app, enable the built-in MCP server from Data Peek settings, then connect the MCP client to the local server. Optional CLI companion: npm install -g @data-peek/cli.
Author
Rohith Gilla
Publisher
Independent Maintainer
Last updated
Oct 7, 2026
Quick brief

Use Data Peek when an operator wants an agent to inspect live database state without handing the model raw credentials or an unrestricted SQL shell. The workflow is to install the Data Peek desktop app, add the PostgreSQL, MySQL, SQL Server, SQLite, or ClickHouse connection, enable the built-in MCP server, and connect an MCP-capable agent to the exposed tools. Data Peek keeps credentials in the operating system keychain, caps agent reads, runs read-only polling for watch-style access, and requires the human operator to approve writes in the app before changes are committed.

How it works

What this skill actually does

Invoke this instead of using the database client normally when the agent needs schema context, query results, diagnostics, or approval-gated write assistance as part of a repeatable investigation or maintenance workflow. The boundary is the guarded MCP handoff around existing Data Peek connections: the agent can ask for structured database context and proposed SQL, while the human remains responsible for connection setup, write approval, and reviewing generated outputs. This is not a generic SQL client listing; publish it as an MCP database-access workflow for supervised agent operations.