material_fetch

Automated keyword-driven material discovery uses template-generated queries to find sources, applies hard screening via filters, and locks vetted results into a persistent memory pool for downstream tasks.

Scheduled content ingestion runs periodically to fetch materials, then applies deterministic hard screening through filters.yaml to discard low-quality or irrelevant items before they reach review.

LLM-based quality review assigns fit scores to fetched materials during a secondary pass, ensuring only high-quality, vetted results are locked and persisted into the workspace memory.

Keyword templates can generate curated queries specifically for life-simulation workspaces, periodically fetching and seeding memory with high-quality sources for storytelling or reasoning.

Template-driven ingestion requires preparing keyword templates and enabling a scheduled trigger, which then executes deterministic screening via filters.yaml to discard irrelevant materials automatically.

Scheduled material ingestion includes state tracking and failure handling to ensure robust lifecycle management, preventing fetch cycle disruptions and maintaining consistent memory pool population.