receipt-per-edge

What it does

Attaches the source document, the extraction run ID, and the schema version to every edge at the moment it's written, so provenance travels with the claim from the start instead of being reconstructed after the fact (the write-path habit from Step 8).

Inputs

• A newly created edge.
• The metadata of the run that produced it: source document, run ID, schema version in effect.

Outputs

• The same edge, now carrying three provenance fields alongside its subject/predicate/object.

Failure mode if skipped

A claim sits in the graph with no way to trace which document, which extraction run, or which schema version put it there — so nothing can check it, and nothing can un-trust it if the source turns out unreliable.

Link to starter kit

Kit: starters/receipt-per-edge/README.md