Skip to content

JPeetz/agent-skills / skills/data/dbt-data-transformation

dbt-data-transformation

Verified

Use this skill when working with dbt (data build tool): model development, testing, documentation, semantic layer (MetricFlow), dbt Mesh governance, incremental strategies, and migration workflows. Follows dbt Labs best practices for data transformation and multi-project collaboration.

FileKindSize
CHANGELOG.mdmarkdown2.7 KB
LICENSEother1.1 KB
SKILL.mdskill md44.6 KB
evals/eval_cases.jsonconfig3.9 KB
evals/evals.jsonconfig3.7 KB
references/dbt-mesh-governance.mdmarkdown15.6 KB
references/dbt-model-patterns.mdmarkdown17.1 KB
references/dbt-semantic-layer.mdmarkdown21.0 KB
scripts/dbt-test-runner.shexecutablescript11.0 KB
scripts/validate-dbt-project.shexecutablescript9.0 KB

evals/evals.json

View on GitHub
[
  {
    "id": "dbt-001",
    "description": "Create a dbt model with staging/marts structure",
    "input": "Create a dbt model that transforms raw orders into a customer_orders mart.",
    "shouldTrigger": true,
    "expect": "Agent routes to Level 1 (Analytics Engineering): establishes source/staging models (stg_ prefix, views) feeding a marts model (mart or fct/dim prefix, table materialization) with ref() for dependencies.",
    "assertions": [
      "staging models use stg_<source>__ prefix and view materialization",
      "marts models use fct_/dim_ naming with table materialization",
      "model dependencies use ref(), never hardcoded table names"
    ]
  },
  {
    "id": "dbt-002",
    "description": "Write data tests for staging models",
    "input": "Write data tests for my staging models to ensure referential integrity.",
    "shouldTrigger": true,
    "expect": "Agent routes to Level 2 (Testing): generic tests in .yml (unique, not_null, relationships) plus custom generic/singular tests with severity thresholds where needed.",
    "assertions": [
      "generic tests (unique, not_null, relationships) are declared in .yml",
      "referential integrity is covered via relationships tests",
      "severity thresholds are set where tests may not be blocking"
    ]
  },
  {
    "id": "dbt-003",
    "description": "Semantic layer with MetricFlow",
    "input": "Define MetricFlow semantic models and metrics for revenue reporting.",
    "shouldTrigger": true,
    "expect": "Agent routes to Level 3 (Semantic Layer): defines semantic models, measures, dimensions, and metrics, then exposes a saved query for reporting.",
    "assertions": [
      "semantic models declare measures and dimensions",
      "metrics are defined on top of semantic models",
      "a saved query is created for the reporting use case"
    ]
  },
  {
    "id": "dbt-004",
    "description": "dbt Mesh cross-project refs",
    "input": "Set up cross-project refs between our finance and marketing dbt projects.",
    "shouldTrigger": true,
    "expect": "Agent routes to Level 4 (dbt Mesh): exposes models via public grouping/access controls and wires cross-project ref() with model contracts.",
    "assertions": [
      "models are marked public with an access group",
      "model contracts enforce the interface between projects",
      "cross-project ref() is used instead of duplicating logic"
    ]
  },
  {
    "id": "dbt-005",
    "description": "Debug a failed incremental build",
    "input": "Troubleshoot this failed incremental model build on our daily dbt job.",
    "shouldTrigger": true,
    "expect": "Agent routes to Level 5 (Platform & Ops) and follows the debugging workflow: dbt parse → compile → run --select the model → inspect compiled SQL → test upstream → dbt --debug run.",
    "assertions": [
      "debugging workflow starts with dbt parse and compile",
      "the failing model is isolated with dbt run --select",
      "compiled SQL is inspected before code changes"
    ]
  },
  {
    "id": "dbt-006",
    "description": "Near-miss negative: general SQL writing",
    "input": "How do I write a SELECT statement?",
    "shouldTrigger": false,
    "expect": "General SQL questions are an explicit do-not-trigger case routing to query-expert skills.",
    "assertions": [
      "skill does not trigger on general SQL questions"
    ]
  },
  {
    "id": "dbt-007",
    "description": "Near-miss negative: orchestration and platform admin",
    "input": "How do I schedule an Airflow DAG?",
    "shouldTrigger": false,
    "expect": "Data pipeline orchestration is an explicit do-not-trigger case for pipeline/orchestration skills.",
    "assertions": [
      "skill does not trigger on orchestration questions"
    ]
  }
]
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall