Identifies data quality issues and pattern variations to guide extraction design.
npx skills add https://github.com/memect/kc --skill data-sensibility
Please help me install this Agent Skill. Skill: data-sensibility Source: https://github.com/memect/kc/tree/main/template/skills/zh/data-sensibility Command: npx skills add https://github.com/memect/kc --skill data-sensibility
Build intuition about document data before extraction design. Use before designing any extraction schema or regex pattern, when onboarding a new document type, or when extraction accuracy is unexpectedly low and you suspect a data assumption is wrong. It covers systematic observation of raw documents, spot-checking extracted results, distribution analysis, and recognizing suspicious patterns. If you are about to write code that touches document data and you have not read at least five documents end-to-end, stop and use this skill first.
Read 3-5 complete documents of a new type to surface data patterns before designing any extraction logic.
High-intent search queries and answers about installing and using this skill.
Unexpected document extraction accuracy drops often stem from broken data assumptions. Systematically observe raw documents and spot-check extracted values to detect anomalous fields and pattern mismatches before adjusting your extraction logic.
To onboard a new document type, read 3-5 representative documents end-to-end first. This systematic data observation surfaces stable anchors and field variability, directly informing your extraction schema design and reducing errors.
Data sensibility is the process of building intuition about document data quality and pattern variations before coding extraction logic. You need it when onboarding new document types, validating data assumptions, or assessing data consistency across samples.
Check data quality by spot-checking extracted values and running distribution analysis on your samples. This detects pattern mismatches and identifies suspicious data anomalies before you commit to specific extraction rules or regex patterns.
If you suspect your data assumptions are wrong, stop coding and read at least five documents end-to-end. Systematic observation of raw documents helps identify data quality issues and pattern variations to correct your extraction design.
Designing extraction rules without reading raw documents risks missing data quality issues and pattern variations. Systematic observation of 3-5 complete documents is required to surface stable anchors and field variability for accurate feature design.
raphaelmansExtract reusable design tokens, components, and patterns into a design system.
jsagirExtract structured intelligence from meeting transcript text using regex patterns.
Ameen-AlamExtract PDF text and tables, fill forms, and annotate fields.
kevinlinProfile datasets to generate summaries, null counts, and quality assessments.
