Questions that come up on day three of using any security tool.
cairnlock:ignore RULEID on the offending line, add the rule
id to disabled_rules in cairnlock.toml, or set
--fail-on never for one run. If the pattern itself is wrong,
please open a false-positive report using the issue template — that fixes
it for everyone.low, which fails on anything.
Teams adopting on a large repo usually start with
--fail-on high, record a baseline with
--write-baseline, and tighten over time.core covers agent context
files with agent-specific rules. extended swaps in the
generic rules and also widens discovery to source files (Python,
JavaScript, TypeScript) and dependency files, because those rules were
written for code. See the rule reference for
the exact split.verify --base. It reads the lockfile as it
exists at the base ref with git show, so a rehash in the
same change is its own finding and fails the run. Add CODEOWNERS on
cairnlock.lock so a human also reviews that file.0 clean, 1 findings or changes,
2 usage or environment error — bad ref, missing base
lockfile, shallow clone. Scripts can rely on 1 meaning "look at this"
and 2 meaning "fix how you called it".cairnlock init again to rebuild it from the current
tree, then review the diff before committing. If the old lockfile is in
git history, git show HEAD:cairnlock.lock shows what was
tracked, and git checkout HEAD -- cairnlock.lock restores
it exactly.scan --write-baseline to record today's findings,
commit the baseline, and gate on new findings only. Fix the backlog at
your own pace; new problems still fail CI.scan --format sarif --output cairnlock.sarif
produces SARIF 2.1.0 with rule help URIs and partial fingerprints, and
the README shows the upload step. The composite action does the verify
and scan for you.