data-analysis-validator

Audit analytical reports by reproducing calculations, verifying methods, and rating decision readiness.

Updated Jan 28, 2026
One-click install
npx skills add https://github.com/scanady/nexus-skills --skill data-analysis-validator-scanady
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: data-analysis-validator
Source: https://github.com/scanady/nexus-skills/tree/main/skills/data-analysis-validator
Command: npx skills add https://github.com/scanady/nexus-skills --skill data-analysis-validator-scanady

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Stakeholders often share dashboards, notebooks, and reports whose numbers, methods, or charts cannot support the decisions they drive. This Skill audits an existing analysis end to end, tracing claims through data, calculations, and visuals before anyone relies on them. ## Core Features & Use Cases - Claim Verification: Recompute headline metrics from raw numerators and denominators, and check filters, joins, aggregation grain, and denominators for silent errors. - Method and Causal Review: Challenge causal language, selection bias, survivorship bias, partial-period comparisons, and denominator drift using a risk-weighted review playbook. - Visual Integrity Checks: Inspect final rendered charts for misleading axes, scales, labels, and titles that overstate the evidence. - Use Case: Before circulating a quarterly retention dashboard to executives, ask for a validation review to receive a report classifying findings as blockers, caveats, or improvements, with an overall rating of ready to share, share with caveats, or needs revision. ## Quick Start Review this analysis and tell me whether its conclusions are supported well enough to share with stakeholders.

Frequently Asked Questions about data-analysis-validator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I validate a data analysis before sharing it with stakeholders?▼

Provide the report, notebook, spreadsheet, SQL, or dashboard along with the decision it supports. The review recomputes headline numbers, checks definitions, filters, joins, and visuals, then issues a readiness rating with classified findings and required caveats.

How to check if a dashboard's numbers are correct?▼

Headline metrics are independently recomputed from raw numerators and denominators rather than reverse-engineering rounded outputs. Row counts, join behavior, aggregation grain, and comparison periods are verified, with evidence locations recorded for each check.

What common errors does an analysis review look for?▼

The review targets join explosions, survivorship bias, partial-period comparisons, denominator drift, averages of averages, timezone mismatches, and selection on outcome. It also flags unsupported causal language and misleading chart design.

Can this review a raw dataset or build a new model?▼

No. It audits an existing analytical artifact or concrete claim, not raw dataset profiling, data cleaning, or model fitting. For those tasks, companion skills for business context, statistical modeling, or pipeline architecture are recommended.

What does the validation report output include?▼

The report contains an overall assessment (ready to share, share with caveats, or needs revision), decision-critical findings classified by impact, reproduction checks with evidence locations, method and visual reviews, stakeholder caveats, and unrun checks.