supervisor-api
Replace the client-side agent loop with Databricks Supervisor API (hosted tools + client-side function tools). Use when: (1) User asks about Supervisor API, (2) User wants Databricks to run the agent loop server-side, (3) Connecting Genie spaces, UC functions, agent endpoints, or MCP servers as hosted tools, (4) Mixing client-side function tools (Python callables your app executes) with hosted tools.
75
92%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Advisory
Suggest reviewing before use
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that clearly defines a specific technical capability (Databricks Supervisor API integration), lists concrete actions, and provides explicit trigger conditions via a well-structured 'Use when' clause with four distinct scenarios. The description is concise yet comprehensive, using domain-specific terminology that users would naturally employ when seeking this functionality.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: replacing client-side agent loop, using Databricks Supervisor API, connecting Genie spaces/UC functions/agent endpoints/MCP servers as hosted tools, and mixing client-side function tools with hosted tools. | 3 / 3 |
Completeness | Clearly answers both 'what' (replace client-side agent loop with Databricks Supervisor API) and 'when' with an explicit 'Use when:' clause listing four specific trigger scenarios. | 3 / 3 |
Trigger Term Quality | Includes highly specific natural keywords users would say: 'Supervisor API', 'Databricks', 'agent loop', 'server-side', 'Genie spaces', 'UC functions', 'agent endpoints', 'MCP servers', 'hosted tools', 'function tools'. These are the exact terms a developer working with Databricks would use. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with a clear niche around Databricks Supervisor API specifically. The combination of Databricks-specific terminology (Genie spaces, UC functions) and the specific architectural pattern (replacing client-side agent loop with server-side) makes it very unlikely to conflict with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a strong, well-structured skill that provides a complete, actionable guide for adopting the Databricks Supervisor API. Its main strengths are the executable code examples, clear step-by-step workflow, comprehensive permission tables, and good progressive disclosure to related skills. The primary weakness is some redundancy in the tracing code (appearing in both the main agent and the dedicated tracing section) and occasional verbosity in explanatory sections that could be tightened.
Suggestions
Remove the duplicated tracing code — the full agent.py in Step 3 already includes _get_trace_destination and _extra_body, so the 'Enabling Tracing > Setup' section could simply reference Step 3's code rather than repeating a longer variant with different error handling.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient with good code examples and tables, but it's quite long (~300+ lines) with some redundancy — the tracing setup code appears twice (once in the main agent.py and again in the dedicated Tracing section), and the tracing section includes verbose error messages and explanations that could be trimmed. The MCP approval flow explanation is clear but could be more concise. | 2 / 3 |
Actionability | Highly actionable with fully executable Python code, complete tool declaration examples for every tool type, concrete bash commands for testing/deploying, specific permission tables, and copy-paste ready agent.py code. The OAuth scopes table and permission matrices are particularly useful concrete references. | 3 / 3 |
Workflow Clarity | Clear 5-step sequential workflow (Install → Declare Tools → Update Agent → Grant Permissions → Test/Deploy) with explicit validation via test command. The MCP multi-turn approval flow is clearly sequenced with numbered steps. Authentication options are clearly delineated between OBO and service principal modes with specific guidance for each. | 3 / 3 |
Progressive Disclosure | Effectively delegates detailed topics to other skills (supervisor-api-client-function-calling, supervisor-api-background-mode, add-tools) and external docs (MLflow distributed tracing docs). References are one level deep and clearly signaled. The main content stays focused on the core Supervisor API pattern while pointing elsewhere for specialized flows. | 3 / 3 |
Total | 11 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- databricks/app-templates
- Commit
1c88215
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.