This page is the fastest path to a first successful Limier run.
- Docker available to the current user
- One of these ways to run Limier:
- a built binary
- the published container image
- a local Go toolchain if you want to build from source
- Linux plus
bpftraceif you want host-signal capture
::: warning Host-signal capture is Linux-only
If capture_host_signals is enabled and Limier cannot start that backend, the run becomes inconclusive instead of silently dropping process coverage.
On macOS, Windows, or CI runners without bpftrace, set capture_host_signals: false in your scenario.
:::
Download a release asset from GitHub Releases and place limier somewhere on your PATH.
Then confirm the install:
limier versionRelease tags also publish an OCI image:
ghcr.io/room215/limier:<version>This is a good fit for CI or environments where you do not want to manage a local install.
If you are working from this repository directly:
go build -o ./bin/limier .
./bin/limier versionThe included sample fixture lives in this repository, so the commands in this section assume you are in a local checkout of room215/limier.
That sample compares:
- ecosystem:
npm - package:
left-pad - current version:
1.0.0 - candidate version:
1.1.0 - fixture:
fixtures/npm-app - scenario:
scenarios/npm.yml - rules:
rules/default.yml
The fastest path is still the repository-owned sample runner:
sh ./examples/ci/run-sample.shThat script builds ./bin/limier, runs the review, and renders a build summary.
Run the same sample directly with your installed limier binary:
mkdir -p out/limier
limier run \
--ecosystem npm \
--package left-pad \
--current 1.0.0 \
--candidate 1.1.0 \
--fixture fixtures/npm-app \
--scenario scenarios/npm.yml \
--rules rules/default.yml \
--report out/limier/report.json \
--summary out/limier/summary.md \
--evidence out/limier/evidence
limier render \
--format build-summary \
--input out/limier/report.json \
--output out/limier/build-summary.mdFrom the repository checkout, mount the checkout at the same absolute path inside the container:
Mounting /var/run/docker.sock gives the Limier container control over the host Docker daemon so it can create the fixture containers. Use this only on a trusted local machine or an ephemeral or isolated CI runner.
mkdir -p out/limier
docker run --rm \
--user "$(id -u):$(id -g)" \
-v /var/run/docker.sock:/var/run/docker.sock \
-v "$PWD:$PWD" \
-w "$PWD" \
ghcr.io/room215/limier:<version> \
run \
--ecosystem npm \
--package left-pad \
--current 1.0.0 \
--candidate 1.1.0 \
--fixture fixtures/npm-app \
--scenario scenarios/npm.yml \
--rules rules/default.yml \
--report out/limier/report.json \
--summary out/limier/summary.md \
--evidence out/limier/evidence
docker run --rm \
--user "$(id -u):$(id -g)" \
-v "$PWD:$PWD" \
-w "$PWD" \
ghcr.io/room215/limier:<version> \
render \
--format build-summary \
--input out/limier/report.json \
--output out/limier/build-summary.mdThe sample writes:
out/limier/report.jsonout/limier/summary.mdout/limier/build-summary.mdout/limier/evidence/
After a run:
- Open
summary.mdfor the short answer. - Open
report.jsonif you need the full structured result. - Open
evidence/when you want the raw stdout, stderr, and event capture behind the verdict.
You can also re-explain a completed report without rerunning the fixture:
./bin/limier inspect --input out/limier/report.jsonIf you installed Limier on your PATH, use limier inspect instead. If you are using the container image, run inspect the same way you ran run, with the repository checkout bind-mounted into the container.
Or render the same report again for a downstream surface:
./bin/limier render --format build-summary --input out/limier/report.jsonIf you installed Limier on your PATH, use limier render instead.
Limier gives an operator recommendation:
good_to_go: nothing suspicious enough was found with the current rulesetneeds_review: the change may be legitimate, but a human should inspect itblock: the change matched a hard-block rule and should not be approved yetrerun: the run was inconclusive or unstable
See Understand Results for how to interpret each one.
- Use Review Your Own Project when you are ready to point Limier at a real dependency upgrade.
- Use Scenario File and Rules File when you want to customize how Limier runs and evaluates your fixture.