validating-temporal-fields

What problem does it solve?

This Skill prevents bad temporal data from slipping into incident corpora, vulnerability registries, and other text-derived datasets by catching future-dated records, year-extraction mistakes, and event-versus-disclosure confusion before ingest succeeds.

Core Features & Use Cases

• Future-date rejection: Flags any event or disclosure date that lands after the pipeline's deterministic today value.
• Min-year fallback: Recommends choosing the earliest plausible year from free text instead of the latest year mentioned, avoiding max-year leakage from contextual references.
• Separate temporal fields: Keeps event_date and disclosure_date distinct so cross-field validation can detect impossible combinations.
• Use case: A data team ingesting an incident database can quarantine suspicious rows, document the failing invariant, and keep CI runs reproducible across days and environments.

Quick Start

Use this skill to validate incident_date, disclosure_date, and text-extracted year fields in your ingest pipeline before accepting the batch.