What is the Databricks Lakehouse MCP skill?
Databricks Lakehouse MCP is a mcp skill by databricks for Claude Code, Cursor, Codex and other AI coding agents. Discover, connect to, or recover the hosted Databricks Lakehouse MCP (system.ai.lakehouse), then load its guidance for inspecting, editing, running, and reading results from Databricks notebooks. Use for notebook work in a Databricks workspace or when the user asks to set up the Lakehouse MCP.
The hosted system.ai.lakehouse service provides notebook tools and the detailed instructions for using them. This local skill handles discovery and connection; load operational guidance from the connected service for the user's task.
When to use it
Use for notebook work in a Databricks workspace or when the user asks to set up the Lakehouse MCP.
What the SKILL.md covers
- Connect or recover
- Retrieve the task guidance
- Author
- databricks
- Pack
- databricks/databricks-agent-skills (30 skills)
- Category
- MCP, #80 of 120
- Alyn Score
- 61 / 99
- Files
- 4, includes references
- License
- Not stated
- Last updated
- 2 days ago
- Works with
- Claude Code, Claude, Cursor, Codex, Gemini CLI, Copilot
Databricks Lakehouse MCP compared to similar skills
| Skill | Alyn Score | GitHub stars | Updated |
|---|---|---|---|
| Databricks Lakehouse MCP · databricks | 61 | 407 | 2 days ago |
| MCP Integration · Anthropic | 85 | 48k | 2 days ago |
| A4a · sickn33 | 76 | 50k | 2 days ago |
| MCP Integration · aiskillstore | 63 | 488 | 2 days ago |
Install Databricks Lakehouse MCP
Pick your agent. Every command is ready to paste.
npx alyn-cli add databricks-lakehouse-mcpDetects your agents (Claude Code, Cursor, Codex and 70+ more) and installs to the right folder. Add -g to install globally.
npx skills add databricks/databricks-agent-skills --skill databricks-lakehouse-mcpOr use the open skills CLI directly.
01
Install it
Run the command above, or upload the folder in Claude.
02
Ask naturally
Describe the task. Your agent loads databricks-lakehouse-mcp when the request matches its description.
03
Name it if needed
If it does not trigger, say “use the databricks-lakehouse-mcp skill”. In Codex, type $databricks-lakehouse-mcp.
Databricks Lakehouse MCP
The hosted system.ai.lakehouse service provides notebook tools and the detailed
instructions for using them. This local skill handles discovery and connection;
load operational guidance from the connected service for the user's task.
Connect or recover
First use the client's tool discovery to look for configured or deferred
Lakehouse MCP tools. If they are available, proceed to skill retrieval.
Otherwise, use the setup instructions shown for system.ai.lakehouse in the
user's Databricks workspace to configure and authenticate the connection in
Claude Code or Codex. Use the service's current setup instructions for connection
details; do not guess an endpoint or installer command.
If system.ai.lakehouse or its setup instructions are unavailable, or preview
access is denied, direct the user to their Databricks account representative to
establish availability and request access. Installing this skill does not grant
MCP access. Missing tools alone do not show whether the connection is
unconfigured or access is unavailable.
For an authentication error, follow the client's sign-in recovery. For an access denial, use the relevant access-request path. For a network or service error, guide retry or diagnosis. If the cause is unclear, keep it unknown. After explaining a blocker, avoid repeating the same setup prompt on later turns; continue through an existing supported path when appropriate for the request.
Use databricks-core when CLI authentication or profile selection is needed.
Retrieve the task guidance
- Discover the connected server's advertised skill tools. Call
list_skills(query="<short task keyword>")to search skill IDs, names, and descriptions by case-insensitive substring. Omitqueryto browse the available guides, including when a search returns no matches. Follow any returnedpage_tokento see more results. - Load the relevant guide with
load_skill(skill="<returned skill id>"). This returns the fullSKILL.mdinstructions and supporting-file metadata. - Read the supporting files needed for the task with
read_skill_file(skill="<returned skill id>", path="<returned file path>"). Use the exact paths returned byload_skillorlist_skill_files. To browse more files, calllist_skill_files(skill="<returned skill id>"); an optionalprefix, such as"references/", narrows the list. Follow any returned next-call instructions to continue browsing with the same filters andpage_token. - Follow the guide and current tool descriptions to perform the task. Load more guides when the task requires them, and reuse applicable guidance while it remains in context. Refresh the relevant guidance after reconnecting or detecting a change in the service's capabilities.
Skill retrieval is a useful connection check; no separate healthcheck is needed. It verifies access to that read, not permission to edit a notebook or readiness to run compute. Handle those checks and failures through the relevant operations. If the server does not advertise skill retrieval, report the missing capability and use the service's current setup/recovery guidance.
Tips for Databricks Lakehouse MCP
- • Start a fresh session after installing so your agent picks up the new skill.
- • It bundles reference files. The agent reads them only when needed, so they cost no context until then.
- • It is part of a pack with 30 skills. Install only the ones you need with --skill to keep triggers precise.
- • If two skills overlap, mention the one you want by name.
Questions about Databricks Lakehouse MCP
What does the Databricks Lakehouse MCP skill do?
Discover, connect to, or recover the hosted Databricks Lakehouse MCP (system.ai.lakehouse), then load its guidance for inspecting, editing, running, and reading results from Databricks notebooks. Use for notebook work in a Databricks workspace or when the user asks to set up the Lakehouse MCP.
How do I install Databricks Lakehouse MCP?
Run "npx alyn-cli add databricks-lakehouse-mcp" in your project. The Alyn CLI installs it for Claude Code, Cursor, Codex and 70+ other agents. In Claude on the web, upload the skill folder as a ZIP under Customize > Skills.
Does Databricks Lakehouse MCP work in Cursor and Codex?
Yes. Databricks Lakehouse MCP uses the open SKILL.md format, so it works in Claude Code, Claude, Cursor, OpenAI Codex, Gemini CLI, GitHub Copilot and other agents that support agent skills.
When should I use the Databricks Lakehouse MCP skill?
Use for notebook work in a Databricks workspace or when the user asks to set up the Lakehouse MCP.
What is a good alternative to Databricks Lakehouse MCP?
MCP Integration by Anthropic is the highest-ranked similar skill from another creator, with an Alyn Score of 85 compared to 61 for Databricks Lakehouse MCP.
Is Databricks Lakehouse MCP free?
Databricks Lakehouse MCP is open source on GitHub. It is free to install; check the license before commercial use.
Who made Databricks Lakehouse MCP?
Databricks Lakehouse MCP is published by databricks in the databricks/databricks-agent-skills repository, together with 29 other skills.


