Search results include state-derived filters and, when the index is enabled, Chromadb data.

Key fields:

• document_id — persistent identifier stable across org/watch/mv/undo flows.
• score — similarity score for semantic results (optional).
• snippet — excerpt from the indexed document payload (optional).

Index lifecycle is managed per collection under .dorgy/chroma with a manifest at .dorgy/search.json.