create-tools

Create Databricks tool resources for agents to execute MCP-backed operations.

Updated May 10, 2026
One-click install
npx skills add https://github.com/keqingli1129/agent-langgraph-one --skill create-tools-keqingli1129
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-tools
Source: https://github.com/keqingli1129/agent-langgraph-one/tree/main/.claude/skills/create-tools
Command: npx skills add https://github.com/keqingli1129/agent-langgraph-one --skill create-tools-keqingli1129

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill removes the friction of provisioning the Databricks resources your agent needs before it can call tools, so you can move from idea to working agent faster.

Core Features & Use Cases

  • Guides resource selection: Choose the correct Databricks tool resource (Genie space, Vector Search index, UC function, UC connection, or inline local Python tools) based on what the agent must do.
  • Supports first-time provisioning: Handles cases where the resource does not exist yet, before you wire it into the agent.
  • Sets up the right workflow: Ensures you follow the end-to-end flow: discover existing resources, create missing ones, then add tools and deploy with the required permissions.

Quick Start

Ask your agent to create the Databricks resources needed for the tool access you want (for example, "set up a Genie space so I can query Unity Catalog tables, then tell me what to create next before adding tools").

Frequently Asked Questions about create-tools

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I provision Databricks resources for an AI agent before wiring tools?

Provision Databricks resources for an agent by first discovering existing infrastructure, then creating missing Genie spaces, Vector Search indexes, or UC functions before wiring tools and deploying with required permissions.

What is the correct workflow for setting up Genie spaces and Vector Search indexes?

The correct setup workflow for Genie spaces and Vector Search indexes is discovering existing resources first, provisioning any missing infrastructure, then enabling follow-up steps for tool wiring and permission grants.

How do I create Unity Catalog functions for agent tool access?

Create Unity Catalog functions for agent tool access by provisioning them as required Databricks tool resources, enabling the agent to connect and execute MCP-backed operations for structured data and custom logic.

Can I define inline local Python tools when no external Databricks resource is required?

Yes, you can define inline local Python tools when no external resource is required, serving as a valid alternative to provisioning UC connections for external MCP servers or other Databricks infrastructure.

Does this agent setup workflow support external MCP servers?

Yes, agent setup supports external MCP servers by provisioning UC connections, enabling agents to connect and execute MCP-backed operations for structured data, RAG, and custom logic within the Databricks environment.

Why does my agent setup require discovering existing Databricks resources first?

Agent setup requires discovering existing Databricks resources first to avoid duplicate infrastructure, ensuring you only provision missing Genie spaces or Vector Search indexes before proceeding to tool wiring and permission grants.