Name: supervisor-api
Rating: 72.8 (1 reviews)
Author: databricks

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.

Quality

—

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Advisory

Suggest reviewing before use

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with executable code and a well-sequenced workflow including troubleshooting feedback loops. Its main weakness is redundancy: the tracing section repeats a helper already shown earlier and explains concepts Claude can infer, and some detail-rich sections are inline rather than split into one-level-deep references.

Suggestions

Deduplicate the tracing helper: Step 3 and "Enabling Tracing" both define `_get_trace_destination` — keep one canonical version and have the tracing section reference it instead of re-listing the full function.

Trim the explanatory prose in "Enabling Tracing" (what distributed tracing is, the 4-point "How It Works") to the essential wiring, since Claude can infer the concepts from the headers and code.

Move the full tracing setup and the MCP multi-turn approval flow into short referenced files (one level deep) so the main SKILL.md stays an overview, since those are advanced/optional paths.

Dimension	Reasoning	Score
Conciseness	The body is largely actionable but the "Enabling Tracing" section duplicates the `_get_trace_destination` helper already shown in Step 3 and restates distributed-tracing concepts, padding the file with content that could be tightened or deferred to a reference. It earns the 'mostly efficient but could be tightened' anchor rather than the lean top level.	2 / 3
Actionability	It provides fully executable, copy-paste-ready guidance: a concrete TOOLS list, complete invoke/stream handlers, version-pinned dependency lines, exact `oauth_scopes` YAML, and tabular permission mappings per tool type — all concrete code and commands with no pseudocode.	3 / 3
Workflow Clarity	Steps are clearly sequenced (Step 1 install → Step 5 deploy) with explicit decision branches (OBO vs service principal), and the Troubleshooting section gives error-message-to-fix feedback loops for the API's failure modes, matching the top anchor for clear sequencing with recovery guidance.	3 / 3
Progressive Disclosure	It signals deeper detail via peer skills ("See the supervisor-api-client-function-calling skill", "See the supervisor-api-background-mode skill") but no bundle files exist in this skill's own directory, and the ~80-line inline tracing section is content that could be split into a one-level-deep reference. Structure is present but not optimally separated, fitting the mid anchor.	2 / 3
	Total	10 / 12 Passed

Description

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.

The description is strong across all four dimensions: it names concrete actions, includes explicit and varied trigger terms, answers both what and when, and occupies a distinctive niche. It follows the third-person voice and avoids padding, closely matching the rubric's good examples.

Dimension	Reasoning	Score
Specificity	The description lists multiple concrete actions: "Replace the client-side agent loop with Databricks Supervisor API", "Connecting Genie spaces, UC functions, agent endpoints, or MCP servers", and "Mixing client-side function tools (Python callables your app executes) with hosted tools." It is not vague; it names the specific integration targets and the architectural replacement being performed.	3 / 3
Completeness	It explicitly answers both what ("Replace the client-side agent loop with Databricks Supervisor API (hosted tools + client-side function tools)") and when via an explicit "Use when: (1)...(4)" trigger clause, matching the rubric's top anchor.	3 / 3
Trigger Term Quality	It surfaces natural terms a user would say — "User asks about Supervisor API", "User wants Databricks to run the agent loop server-side", plus the concrete resource names (Genie spaces, UC functions, MCP servers). Coverage of likely user phrasings is strong rather than jargon-only.	3 / 3
Distinctiveness Conflict Risk	The trigger is anchored to the specific, distinctive product "Databricks Supervisor API" with named sub-resources, giving it a clear niche that is unlikely to fire for unrelated skills.	3 / 3
	Total	12 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: databricks/app-templates
Commit: 2a4c792

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.