Filtering Findings
Focus a scan on the findings that matter — by severity, confidence, or both, using floors or exact values.
Chainvet filters findings on two independent axes — severity and confidence — and each axis offers two mutually-exclusive styles: a floor or an exact set.
Floor: at or above
chainvet scan -s high contracts/ # severity >= high
chainvet scan -c high contracts/ # confidence >= high (most precise only)
chainvet scan -s medium -c medium . # combine the two axes-s/--min-severity— report findings at or above this severity.-c/--min-confidence— report findings at or above this confidence.
Exact: only these values (repeatable)
chainvet scan --severity medium contracts/ # exactly medium
chainvet scan --severity medium --severity high contracts/ # medium OR high
chainvet scan --confidence high contracts/ # exactly high confidence--severity— report findings of exactly this severity (repeat to allow several).--confidence— report findings of exactly this confidence (repeatable).
The one rule
Floor and exact conflict — per axis
You cannot combine -s with --severity (nor -c with --confidence) — the
floor and the exact filter conflict on the same axis and the CLI will reject it.
The two axes are independent, so -s high --confidence high is fine.
Severity vs. confidence
The two axes answer different questions:
-s high— show me the highest-impact findings.-c high— show me only the findings the analyzer is most confident about (the most precise detections).
See Severity & Confidence for the model.
Filters are display-only
JSON stays complete
Filters affect only the human-readable pretty output. The -f json payload is
always complete and unfiltered, so downstream tooling gets every finding regardless
of the flags you pass.