abd-ulbasit/upgradescope gate

Gate rendered Kubernetes manifests on upgrade readiness: removed and deprecated APIs for a target version, as SARIF, outputs and a step summary. Exit code 2 (findings at/above fail-on, or an unknown verdict unless allow-incomplete) fails the step.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
pathFile or directory of rendered manifests to scan (e.g. the output of helm template / kustomize build).yes—
targetTarget Kubernetes minor version, e.g. "1.36".yes—
fail-onFail the gate when findings at/above this severity exist: blocker | warning | never.noblocker
allow-incomplete"true" passes scan --allow-incomplete: the gate then fails on findings alone, not on an unknown verdict (a required check, such as kb-coverage for a target past the knowledge base's horizon, was not assessed). The verdict output still says unknown. "true" or "false"; needs upgradescope v0.2.0 or later.nofalse
versionWhich upgradescope to run: a release tag such as v0.2.0, "latest" (the newest stable release; GitHub's latest skips prereleases), or "preinstalled" (the one already on PATH). Left unset, it is the release of the action ref: at a release tag (@vX.Y.Z or @vX.Y.Z-rc.N), that release; at a full commit SHA, the release tag that points at the commit (git ls-remote), or latest with a warning when no release tag does or the lookup fails; at any other ref (a branch, @v0), or used from another action, latest. For a reproducible gate, pin the action ref (@vX.Y.Z or a commit SHA) and a tag here: @v0 and "latest" move with each release. The archive is verified against the release's checksums.txt; without one for this runner, it is built with `go install` when Go is set up.no""
configPath to an .upgradescope.yaml with ignore rules (scan --config). Suppressed findings are not scored and do not fail the gate; the step summary lists each with its reason. Unset, scan looks for .upgradescope.yaml in path, then at the repository root.no""
baselinePath to the JSON report of an earlier scan (the report-json output, or a write-baseline file; scan --baseline). The gate then fails only on findings that are new since; score and verdict still count all.no""
write-baselineAlso write this scan's JSON report (after suppression) to this path, for a later baseline (scan --write-baseline).no""
namedescription
sarif-filePath to the SARIF report (upload with github/codeql-action/upload-sarif; complete even when the gate fails; not set when the scan itself failed).
report-jsonPath to the JSON report (upgradescope scan --output json).
verdictready | blocked | unknown (no blockers found, but a required check could not run).
ready"true" when the verdict is ready, else "false".
scoreReadiness score, 0-100.
blockersNumber of blocker findings.
warningsNumber of warning findings.
new-blockersBlocker findings not in the baseline, the ones the gate counts (equals blockers without a baseline).
new-warningsWarning findings not in the baseline (equals warnings without a baseline).
suppressedNumber of findings suppressed by ignore rules or annotations (not counted in blockers or warnings).
summary-filePath to the Markdown summary, the same text as the step summary (e.g. for a pull request comment).