artifact-query

To search past engineering handoffs for historical precedents, query a SQLite FTS5 artifact index using semantic full-text search. The system applies BM25 ranking to return the most relevant plans and continuity records.

You can filter artifact index search results by artifact type and outcome using structured query parameters. A planning mode is available to specifically separate successful and failed handoffs.

Yes, you need an initialized SQLite FTS5 index to perform semantic full-text search on continuity records. This database backend enables the BM25 ranking required to surface relevant handoffs.

Yes, you can get historical plan records returned as structured JSON from a full-text search by enabling optional JSON output. You can also apply limits to control the number of artifacts returned.

The best way to avoid repeating known engineering mistakes during project planning is to search an artifact index for prior handoffs and continuity records. This surfaces past failures and successes using BM25-ranked semantic search.

You should use BM25 semantic search for debugging and implementation tasks when you need to cite what worked, what failed, and which files to reference from similar past implementations before starting work.