DAG Diagnosis (Airflow)

What it does

This skill equips an agent with a step-by-step Airflow DAG diagnosis workflow: identify failing runs, fetch task logs, categorize failure types (data, code, infra, dependency), assess impact, propose immediate fixes and long-term prevention measures, and provide ready-to-run commands. It is focused on pragmatic, operator-friendly remediation rather than abstract guidance.

When to use it

Use this skill when you need a thorough root-cause analysis of failed Airflow DAG runs — for example: "diagnose and fix the pipeline", "why is this DAG failing", or when a task consistently errors after a deploy. It is appropriate for both single-run triage and recurring failure investigations.

What's included

• Scripts: none embedded (has_scripts=false)
• References: no external references file (has_references=false)
• Instructions: clear CLI-first procedure using the provided af wrapper (uvx) with commands for logs, runs, health checks, and run management; explicit output structure (Root Cause, Impact, Immediate Fix, Prevention, Quick Commands).

Compatible agents

Best used by agents with CLI and observability access (Cursor/Copilot-style tooling or Claude Code with shell access) that can run af commands and inspect logs.