asmuelle/cargo-impact

Blast-radius analysis for Rust PRs — SARIF code scanning + sticky PR comment

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Activelast commit Apr 23, 2026
License
Apache 2.0

Pinned Snippet

workflow.ymlSHA-pinned
uses: asmuelle/cargo-impact-action@c857cefe7ecc8be47527ea3615e82872e42b0ff2 # v1.0.0

tags can be moved; commit SHAs can't. why a SHA?

namedescriptionrequireddefault
versioncargo-impact version to install. `"latest"` pulls whatever `cargo install cargo-impact` resolves today; a specific version (e.g. `"0.3.0"`) is pinned for reproducibility. nolatest
sinceGit ref to diff against. Defaults to the PR base SHA on `pull_request` events, or `HEAD~1` otherwise. Override when you need a fixed reference (release branch comparisons, etc.). no""
fail-onSeverity gate: `"high"`, `"medium"`, `"low"`, or `""` for no gate. Empty by default — users can start informational-only and tighten once the signal is trusted. no""
confidence-minHide findings below this confidence (0.0-1.0). Default 0.0 (show all). no0.0
featuresComma-separated Cargo features to activate, as passed to `cargo impact --features`. Empty uses the manifest's default set. no""
all-featuresActivate every feature in the manifest. Mutually exclusive with `no-default-features` per cargo-impact's flag contract. nofalse
no-default-featuresSkip the manifest's default feature list.nofalse
semver-checksRun `cargo-semver-checks` (slow, requires the tool to be installable on the runner). Off by default. nofalse
upload-sarifUpload the SARIF file to GitHub code scanning. Requires the workflow to have `security-events: write` permission. notrue
comment-on-prPost the markdown report as a sticky PR comment. Requires the workflow to have `pull-requests: write` permission and only runs on `pull_request` events. notrue
namedescription
sarif-pathPath to the generated SARIF file (always produced).
comment-pathPath to the markdown PR-comment body (always produced).
exit-codecargo-impact's exit code from the gate run. 0 = clean, 1 = `fail-on` tripped, 2 = tool error. Useful for conditional downstream steps.