Data Ops Review¶
ContributorsMyles Henaghan
When to use: Deliberately invoke in Agent mode against a Databricks Asset Bundle monorepo (or similar multi-bundle ETL repo) to produce a point-in-time maturity assessment. Re-run periodically (e.g. monthly, or after a sprint focused on quality improvements).
Instructions¶
- Read references/dataops-review-prompt.md before inspecting the target repo.
- Keep test automation and data quality as separate dimensions — do not conflate them.
- Produce a structured markdown report with an At a glance block first.
- If asked to save output, write
docs/dataops-quality-review-YYYY-MM.mdin the target repo and compare against any prior review under Progress since last review. - If the user agrees, use the richest canvas-like output available in the
current environment. If that capability is unavailable, generate a regular
.mdreport with the same sections and evidence. - For shape and depth of a finished report, see the redacted use case assess-dataops-maturity. Do not duplicate the skill procedure there — that file is sample outcome only.
What this review covers¶
| View | What it answers |
|---|---|
| Repository overview | Bundle count and layout, shared config, CI/CD, job definitions |
| Practice adoption | % of bundles with config, PR validation, lint, tests, manifests, DQ, schema enforcement |
| Top opportunities | Highest-impact gaps ranked by value vs effort |
References¶
Troubleshooting¶
| Symptom | Likely cause | Fix |
|---|---|---|
| Conflated test vs DQ findings | Treating pytest and table expectations as one bucket | Re-read the test automation vs data quality table in the prompt |
| Vague recommendations | No file/line evidence | Cite concrete paths, counts, and % before recommending |
| Missing progress section | No prior review in target docs/ |
Skip § Progress; note first recorded run |