databricks-solutions

Model Serving Auth — AI skill for Claude Code

AI community

Model Serving auth — system auth vs OBO (App → Supervisor → Serving → Vector Search).

How to install Model Serving Auth

Installs to ~/.claude/skills/databricks-solutions-agents-with-mcp-on-databricks-apps-model-serving-auth/SKILL.md

Terminal

mkdir -p ~/.claude/skills/databricks-solutions-agents-with-mcp-on-databricks-apps-model-serving-auth && curl -fsSL https://raw.githubusercontent.com/databricks-solutions/agents-with-mcp-on-databricks-apps/HEAD/.claude/skills/model-serving-auth.md -o ~/.claude/skills/databricks-solutions-agents-with-mcp-on-databricks-apps-model-serving-auth/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.

What Model Serving Auth does

Model Serving auth — system auth vs OBO (App → Supervisor → Serving → Vector Search)

How a served agent/retriever authenticates to downstream Databricks resources (Vector Search, UC functions, LLM/embedding endpoints). Consult when deploying any custom Model Serving endpoint that calls protected resources — e.g. the Stage-11.2 VS retriever agents. Full analysis: `research/authentication-design-for-model-serving.md`.

The decision: system auth (Option 2) vs end-to-end OBO (Option 1)

The ch

Alternatives in AI

Xerj — Local-first search engine and MCP server for AI agents 12.5k ★
Inference Serving — ai-research-skills vLLM, SGLang, TensorRT-LLM, llama.cpp 5.4k ★
Nocturne Memory — A lightweight, rollbackable, and visual Long-Term Memory Server for MCP Agents 1.3k ★