Find code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.
Exploration provenance
Early exploration treated exact source-range coverage, representation and output budgets as open hypotheses. Downstream coding-task quality replaced those proxy metrics as the deciding evidence: useful starting context can be sufficient even when the caller still needs ordinary reads, and high source coverage does not prove a correct patch.
The completed decision map owns user constraints; the implementation record owns closure and measured limits. Reference research identifies the frozen winner and retained official evidence. Earlier personal-repository pilot records remain local historical material, not portable product validation or supported evaluation entry points.