diff --git a/guides/full-eval.md b/guides/full-eval.md index b630100..07a82c6 100644 --- a/guides/full-eval.md +++ b/guides/full-eval.md @@ -30,6 +30,11 @@ Do not relay agents' reports to the user as they arrive; wait for all of them. Produce ONE report, written to `EVALUATION.md` at the repo root (this file is your only write), then show the user just the Verdict and Priority queue sections. +This path is stable and overwritten in place on each run, so if `EVALUATION.md` is +committed to version control, a subsequent run produces a reviewable diff of how the +evaluation changed over time. Whether to commit it is the user's call — don't commit it +yourself. + EVALUATION.md structure: ```