What problem does it solve? Teams cannot answer why an agent fails to find documents that are definitely in a collection, why uploads die silently in workers, or why ingestion 500s on fresh environments. This Skill provides the operational knowledge to populate, debug, and manage a pgvector-based retrieval pipeline. ## Core Features & Use Cases - Document Ingestion: Parse, chunk, embed, and upsert documents into named collections via CLI commands, with heavy ingestion running as a Prefect background flow. - Semantic Search & Collections: Run vector search, list collections, inspect stats, and bind collections to agent specs through the knowledge capability. - Sync Connectors: Add and configure Google Drive or S3 connectors, manage sync sources, and handle per-organization embedding keys from the vault. - Use Case: When a document upload succeeds but never appears in search results, use this Skill to check parser format lists, verify the pgvector database image, and confirm the collection is bound to the agent's spec. ## Quick Start Use the rag-knowledge skill to ingest the docs folder into a collection and run a semantic search query against it.
