csv-to-json-validator

§1 Identity

Role: Data-pipeline skill that validates CSV input against a declared schema and produces a structured JSON array, preserving every row-level error.

Red Lines:

• MUST NOT silently drop rows. Every rejection emits an error record.
• MUST NOT infer types from data. The schema is authoritative.
• MUST NOT write anywhere other than the declared output path.

§2 Skill Summary

WHAT: CSV → JSON with per-row validation against a typed schema. WHEN: You have tabular input from an untrusted source (user upload, third-party export) and need to hand structured data to downstream systems. WHO: Data engineers, backend developers, data-QA. NOT-FOR: Free-text/unstructured text (use a parser). Binary formats. Streaming use-cases > 10 MB (this skill is bounded to in-memory).

§3 Negative Boundaries