data-critic

Review code for numerical calculation and data logic errors.

1|Updated Mar 2, 2026
One-click install
npx skills add https://github.com/zivtech/harsh-critic --skill data-critic
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: data-critic
Source: https://github.com/zivtech/harsh-critic/tree/main/zivtech-data-skills/critic/.claude/skills/data-critic
Command: npx skills add https://github.com/zivtech/harsh-critic --skill data-critic

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

The data-critic Skill addresses the challenge of identifying numerical errors and data logic flaws in code, ensuring that calculations and data handling are accurate and reliable.

Core Features & Use Cases

  • Formula Verification: Checks every calculation against its specification to ensure accuracy.
  • Assumption Extraction: Identifies and validates assumptions made about data.
  • Fallback & Default Analysis: Audits handling of missing or out-of-range data.
  • Data Provenance Tracing: Tracks data from source to final output to identify transformations and potential issues.
  • Unit Consistency Checking: Ensures correct unit usage and conversions.
  • Statistical Validity Review: Verifies statistical methods and conclusions.
  • Precision & Rounding Audit: Detects and addresses rounding errors and floating-point issues.
  • Multi-Perspective Investigation: Reviews from the perspectives of a data engineer, domain expert, and adversarial input tester.
  • Gap Analysis: Identifies missing validations and error handling.
  • Evidence Requirements: Ensures that findings include specific code references and examples.

Quick Start

Use the data-critic skill to review the 'src/billing/invoice.ts' file for numerical correctness.

Frequently Asked Questions about data-critic

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

FAQPage Schema
How do I verify numerical accuracy and data logic in financial calculations?

Verifying numerical accuracy and data logic in financial calculations requires checking each formula against its specification, auditing fallback handling for missing data, and tracing data provenance from source to output to ensure business rule adherence.

What is the best way to audit floating-point precision and rounding errors in a data pipeline?

Auditing floating-point precision and rounding errors in a data pipeline involves conducting a dedicated precision audit to detect computational issues, verifying unit consistency, and checking statistical validity across all data transformations.

How do you check statistical validity and data assumptions in code?

To check statistical validity and data assumptions in code, extract and validate assumptions made about the data, verify statistical methods and conclusions, and identify missing validations through a comprehensive gap analysis.

Can I review data transformation logic for missing validations and error handling?

You can review data transformation logic for missing validations and error handling by performing a gap analysis and auditing the handling of out-of-range data from the perspective of an adversarial input tester.

Does data logic review work for tracking data provenance in statistical analysis systems?

Data logic review effectively tracks data provenance in statistical analysis systems by thoroughly analyzing each data handling and transformation step to identify potential issues and ensure data integrity.