validate-data · openai — Codex skill: what it does, how to install
Audits analytical reports, SQL queries, metrics, and charts before sharing them with stakeholders or basing decisions on them.
npx skills add openai/plugins --skill validate-data -a codexWhat validate-data does
The skill runs QA on completed analytical artifacts, including spreadsheets, SQL scripts, and notebooks. It verifies that the analysis answers the core business question rather than an easier proxy. It flags hidden exclusions and unsupported causal claims.
During review, it recalculates critical numbers, confirms denominators, and inspects joins for row drops or duplicates. Visuals get checked for broken scales or truncated axes. Where connectors like ~~structured_data or ~~operations_logs exist, the agent queries them to confirm schema details and freshness.
Install this skill before sharing numbers with leadership. If you only need raw table profiling, a data quality companion is a better fit.
What it helps with
- recalculate key metrics and aggregations across SQL or notebooks
- detect hidden filters, sample bias, and flawed cohort logic
- spot misleading visual patterns like cut-off axes or skewed scales
- separate verified findings from speculative recommendations
Example prompts for the agent
- Review the calculations and claims in this notebook before I share it
- Validate the SQL queries and cohort retention metrics in this report
- Audit this presentation for misleading chart scales and unsupported conclusions
How to install validate-data
1Skills CLI (Vercel)
npx skills add openai/plugins --skill validate-data -a codex2Manually: put the files in place
.agents/skills/validate-data/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add openai/plugins --skill validate-data -a claude-code2Manually: put the files in place
~/.claude/skills/validate-data/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/validate-data/
Manually: put the files in place
.cursor/skills/validate-data/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install openai/plugins validate-data2Manually: put the files in place
.github/skills/validate-data/SKILL.mdPlatform CLI
gemini skills install https://github.com/openai/plugins --path plugins/data-analytics/skills/validate-dataManually: put the files in place
.devin/skills/validate-data/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
From the same set201
The whole set →Other versions of validate-data
Similar
Questions about validate-data
How do I install validate-data in Codex?
Run npx skills add openai/plugins --skill validate-data in your terminal. Codex installs the definition to .agents/skills/validate-data/SKILL.md.
What does the validate-data skill do?
It audits data reports, notebooks, and dashboards for calculation errors, deceptive charts, and methodology gaps. It then separates blocking issues from non-critical caveats.
When should I use analyze-data-quality instead of validate-data?
Use validate-data to inspect the business logic, formulas, and conclusions of an analysis. Use analyze-data-quality when the primary problem is raw data hygiene, such as table freshness, missingness, or duplicate rows.