Severity: low.
What happened
1 of 452 rows (a google:gemini-2.5-pro prediction) returned invalid JSON and was dropped (451 scored). The count is surfaced on import (good — issue_count: 1), but recovering it requires re-running the whole job.
Suggested fix
A targeted retry that re-issues only the malformed/failed scenarios (e.g. --retry-malformed on run/import), so a couple of provider hiccups don't change effective N or force a full re-run.
Found while building a digital-twin validation for a 33-respondent SMB survey (smb.csv) end-to-end via zwill guide (survey → import → commit → one-shot marginals → twin-probability job → twin-validate → report). File/line refs are from zwill/ at that checkout.
Severity: low.
What happened
1 of 452 rows (a
google:gemini-2.5-proprediction) returned invalid JSON and was dropped (451 scored). The count is surfaced on import (good —issue_count: 1), but recovering it requires re-running the whole job.Suggested fix
A targeted retry that re-issues only the malformed/failed scenarios (e.g.
--retry-malformedon run/import), so a couple of provider hiccups don't change effective N or force a full re-run.Found while building a digital-twin validation for a 33-respondent SMB survey (
smb.csv) end-to-end viazwill guide(survey → import → commit → one-shot marginals → twin-probability job →twin-validate→ report). File/line refs are fromzwill/at that checkout.