databricks-zerobus-ingest
Build Zerobus Ingest clients for near real-time data ingestion into Databricks Delta tables via gRPC. Use when creating producers that write directly to Unity Catalog tables without a message bus, working with the Zerobus Ingest SDK in Python/Java/Go/TypeScript/Rust, generating Protobuf schemas from UC tables, or implementing stream-based ingestion with ACK handling and retry logic.
71
86%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Risky
Do not use without reviewing
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 narrow, specialized domain (Zerobus Ingest SDK for Databricks Delta table ingestion) with concrete actions and explicit trigger conditions. It uses third-person voice correctly, includes rich domain-specific trigger terms, and has a well-structured 'Use when...' clause covering multiple scenarios. The description is concise yet comprehensive, making it easy for Claude to select this skill precisely when needed.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: building Zerobus Ingest clients, writing to Unity Catalog tables via gRPC, generating Protobuf schemas from UC tables, implementing stream-based ingestion with ACK handling and retry logic. Very detailed and actionable. | 3 / 3 |
Completeness | Clearly answers both 'what' (build Zerobus Ingest clients for near real-time data ingestion into Databricks Delta tables via gRPC) and 'when' with an explicit 'Use when...' clause listing four distinct trigger scenarios. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural terms a user would say: 'Zerobus Ingest', 'Databricks Delta tables', 'gRPC', 'Unity Catalog', 'Protobuf schemas', 'Python/Java/Go/TypeScript/Rust', 'stream-based ingestion', 'ACK handling', 'retry logic'. These are the exact terms a developer working in this domain would use. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with a very clear niche: Zerobus Ingest SDK specifically for Databricks Delta/Unity Catalog ingestion. The combination of Zerobus, gRPC, Protobuf schemas from UC tables, and the specific SDK languages makes it extremely unlikely to conflict with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
72%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured skill with strong progressive disclosure and actionable guidance, including executable code examples and concrete tool usage patterns. Its main weaknesses are redundancy (context reuse and failure handling are duplicated) and missing explicit validation checkpoints after ingestion. The timestamp format fix section feels like a patch note rather than integrated guidance.
Suggestions
Remove the duplicated 'Handling Failures' section — it repeats the same steps already covered in the workflow (step 5) and the 'Context Reuse Pattern' section.
Add an explicit validation step after ingestion (e.g., a SQL query via execute_code to SELECT COUNT(*) from the target table) to confirm records were durably written.
Integrate the 'Critical Learning: Timestamp Format Fix' into the Python example or the protobuf schema guide rather than leaving it as a standalone callout with no code example.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill has some redundancy — the 'Context Reuse Pattern' and 'Handling Failures' sections repeat nearly identical information (error handling steps appear twice), the 'What Is Zerobus Ingest?' section explains concepts that could be trimmed, and the 'Key Concepts' section includes explanations Claude likely already knows (e.g., what gRPC is). The decision table and troubleshooting table are efficient, but overall the document could be tightened significantly. | 2 / 3 |
Actionability | The skill provides a fully executable minimal Python example with real imports and concrete code, specific pip install commands, explicit MCP tool usage patterns with parameter names, and a detailed troubleshooting table with specific solutions. The workflow steps are concrete and reference specific tools and file paths. | 3 / 3 |
Workflow Clarity | The workflow has a clear numbered sequence (steps 0-6) with a retry/fix loop at step 5, but the validation checkpoints are weak — there's no explicit validation step after ingestion to confirm records landed in the Delta table. The prerequisite validation ('never execute without confirming') is mentioned but lacks a concrete verification command. The 'Get schema' step references protobuf generation but doesn't clarify what to do for JSON-only flows. | 2 / 3 |
Progressive Disclosure | Excellent structure with a concise overview, quick decision table routing users to the right sub-file, and two well-organized reference tables pointing to clearly named files (1-setup-and-authentication.md through 5-operations-and-limits.md). References are one level deep and clearly signaled with descriptive 'When to Read' guidance. | 3 / 3 |
Total | 10 / 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-solutions/ai-dev-kit
- Commit
93cb4e3
- 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.