When your dashboard looks off or someone asks if they can trust yesterday's numbers, this checks table timestamps against a freshness scale (fresh under 4 hours, stale up to 24, very stale beyond that). It finds timestamp columns across common ETL patterns, queries for last update times, and reports status in a scannable format. If data is stale, it points you to the relevant Airflow DAG and suggests using the debugging-dags skill to investigate failures. Honestly just saves you from writing the same timestamp queries over and over, and the staleness scale gives you a shared vocabulary when something hasn't updated.
npx -y skills add astronomer/agents --skill checking-freshness --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills