dbt-labs/dbt-agent-skills
A curated collection of Agent Skills for working with dbt, to help AI agents understand and execute dbt workflows more effectively.
91
Does it follow best practices?
Validation for skill structure
migrating-dbt-core-to-fusion
skills/dbt-migration/skills/migrating-dbt-core-to-fusion/SKILL.md
Discovery
75%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 a solid skill description with clear domain specificity and an explicit 'Use when' clause. The main weakness is that it focuses on a single action (classification) rather than listing multiple concrete capabilities, and could benefit from additional trigger term variations that users might naturally use.
Suggestions
Add more natural trigger terms users might say, such as 'dbt errors', 'migration issues', 'Fusion compatibility', or 'what can I fix'
Consider expanding the capabilities beyond classification to include any related actions like 'suggests fixes' or 'generates fix scripts'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (dbt-core to Fusion migration errors) and describes the classification action with specific categories (auto-fixable, guided fixes, needs input, blocked), but doesn't list multiple concrete actions beyond classification. | 2 / 3 |
Completeness | Clearly answers both what (classifies migration errors into actionable categories) and when (explicit 'Use when' clause specifying triaging migration errors to understand fixable vs blocked issues). | 3 / 3 |
Trigger Term Quality | Includes relevant terms like 'dbt-core', 'Fusion', 'migration errors', 'triaging', but missing common variations users might say like 'dbt errors', 'migration issues', 'error categories', or 'fix errors'. | 2 / 3 |
Distinctiveness Conflict Risk | Very specific niche targeting dbt-core to Fusion migration specifically, with distinct terminology unlikely to conflict with general error handling or other migration skills. | 3 / 3 |
Total | 10 / 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 well-crafted skill with excellent workflow clarity and actionability. It provides concrete commands, clear categorization framework, and explicit validation checkpoints throughout the triage process. The main weakness is moderate verbosity—some concepts are repeated across sections and could be consolidated without losing clarity.
Suggestions
Consolidate repeated mentions of 'repro command' and 'repro_command.txt' into a single reference section to reduce redundancy
Combine the 'Important Notes' and 'Anti-Patterns to Avoid' sections since they largely mirror each other
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably efficient but includes some redundant explanations (e.g., repeating 'use the repro command' multiple times, explaining why autofix review matters in multiple places). Some sections could be tightened without losing clarity. | 2 / 3 |
Actionability | Provides concrete, executable commands (dbt debug, dbt compile, uvx autofix command), specific error codes to look for, exact fix patterns (e.g., quote nesting syntax), and clear examples of what to check in output. Copy-paste ready throughout. | 3 / 3 |
Workflow Clarity | Excellent multi-step workflow with explicit validation checkpoints (Step 0 → Step 1 → Step 2), clear sequencing, feedback loops ('After EVERY fix, re-run the repro command'), and explicit handling of cascading errors. The progressive fixing approach includes validation gates. | 3 / 3 |
Progressive Disclosure | Well-structured overview with clear one-level-deep references to detailed materials (error-patterns-reference.md, classification-categories.md). Main skill provides actionable summary while pointing to comprehensive references for deeper detail. | 3 / 3 |
Total | 11 / 12 Passed |
Validation
81%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 9 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
allowed_tools_field | 'allowed-tools' contains unusual tool name(s) | Warning |
metadata_version | 'metadata.version' is missing | Warning |
Total | 9 / 11 Passed | |