What problem does it solve? When a dbt Cloud or dbt platform job fails, error messages are often unclear and failures can be intermittent, making root cause analysis slow and error-prone. This Skill provides a systematic workflow to diagnose failures instead of guessing or blindly modifying tests. ## Core Features & Use Cases - Structured Diagnosis Workflow: Classifies failures into infrastructure, code/compilation, or data/test errors, then applies targeted investigation steps for each type. - Multi-Source Investigation: Uses the dbt MCP Admin API (list_jobs_runs, get_job_run_error), git history review, dbt CLI commands (parse, compile, show), and compiled test SQL to trace root causes. - Resolution Discipline: Enforces creating a fix branch, adding unit or data tests to prevent recurrence, and documenting unresolved findings with an investigation template. - Use Case: A nightly dbt Cloud job fails on a uniqueness test. The Skill pulls the run error via the Admin API, compiles the failing test's SQL, queries the underlying data with dbt show, and identifies that a recent upstream change introduced duplicate keys. ## Quick Start Diagnose why my dbt Cloud job 12345 failed last night and find the root cause.
