What problem does it solve? Maintaining accurate metadata (knowledgeCutoff, family, generation) across roughly 1900 AI model entries in 80 provider files is error-prone: vendors publish cutoffs inconsistently, third-party aggregators copy wrong values, and naming conventions hide family/version info. This Skill provides sourcing rules, derivation regexes, and idempotent codemods to populate these fields correctly at single-model or repo-wide scale. ## Core Features & Use Cases - Authoritative sourcing rules: Defines which sources are acceptable (vendor docs, official Hugging Face cards, tech reports) and which to reject (aggregator sites, AWS Bedrock cards, release-date inference), with per-vendor footguns documented. - Rule-based family/generation derivation: scripts/derive-family.ts maps normalized model ids to family and generation slugs, handling traps like date suffixes, size suffixes, and vendor spelling variants. - Idempotent codemods: scripts/apply-cutoffs.ts and scripts/apply-family.ts insert fields into prettier-formatted model entries without touching existing values, plus scripts/extract-model-ids.ts to enumerate chat-model ids. - Use Case: When a new Claude model PR lands, fill all three metadata fields inline citing the official system card; or run a repo-wide sweep with multi-agent research and adversarial verification to backfill cutoffs across all providers. ## Quick Start Ask the agent to backfill knowledgeCutoff and family metadata for the aiModels provider files using the model-bank-metadata skill's extract, derive, and apply scripts.
