ingest-training-datalake

SUSPICIOUS: the stated purpose mostly matches corpus acquisition, but the skill's core download mechanism is unverifiable and its continuous fetch loop is broader than a minimally scoped planning tool. Risk is driven by undocumented fetcher provenance, autonomous ongoing network actions, and ingestion of arbitrary external content into downstream AI training workflows.