| name | export-sarif |
| user-invocable | false |
| description | Export .kuzushi/findings.json as a SARIF 2.1.0 file (.kuzushi/findings.sarif) so findings are consumable by CI code-scanning, dashboards, and IDEs. Deterministic transform — no analysis. Pass "all" to include reviewed/noise findings too. |
| argument-hint | [all] |
| allowed-tools | Bash |
Export findings as SARIF
Convert the shared findings index into SARIF 2.1.0 for interop with CI / IDE tooling.
Run, using the project working directory as <cwd> (add --all only if $ARGUMENTS is all):
node "${CLAUDE_PLUGIN_ROOT}/scripts/cmd/export-sarif.mjs" --target "<cwd>"
It writes .kuzushi/findings.sarif and prints { sarifPath, resultCount, ruleCount }. By
default only actionable findings are exported (status open/confirmed/proven, or verdict
exploitable/finding); pass all to include reviewed/noise too. Relay the result and the path.
Severity maps to SARIF level (critical/high → error, medium → warning, else → note); each
distinct CWE becomes a SARIF rule; the kuzushi fingerprint is carried as a partial fingerprint.
When NOT to use
- Before any findings exist — run a producer (
/threat-hunt, /taint-analysis, /systems-hunt)
first.
- To analyze or triage — this only reformats existing findings; it makes no security decisions.