Reference
What ci-parity reads, what it reports, and the exact formats it prints. These pages are the contract: rule IDs, messages, config keys, the baseline format and the JSON and SARIF schemas change only with a recorded decision and a new version.
| Page | Covers |
|---|---|
| Configuration | The files scanned, and every .ci-parity.yml key with its default |
| Baseline | .ci-parity-baseline.json and how counts tolerate, lock and ratchet findings |
| Analysis | Steps, shells, command extraction, wrappers, allow comments and invokers |
| Runners | Reading Taskfiles, Makefiles, justfiles and mise tasks, and each runner’s call syntax |
| Rules | P001 to P005, their exact messages, and how suggestions are chosen |
| CLI | check, baseline, fix and version, their flags, and the exit codes |
| Output formats | The text, github, JSON and SARIF formats |
| GitHub Action | action.yml’s inputs, outputs and steps |
| Compatibility | What stays stable between releases, what may change, deprecation, and the platforms tested |
Vocabulary
Section titled “Vocabulary”- Runner: the task runner: Taskfile, make, just or mise.
- Invoker: how CI calls the runner. A shell invoker (
task x,make x,just x,mise run x) or an action invoker (a local composite action such as./.github/actions/run-taskwhose input names the task). - Tool: a command on the tools list. Only tools are flagged; everything else is ignored.
- Finding: one rule hit at one location, with severity
error,warning,toleratedorallowed. - Baseline: the per-workflow count of tolerated P001 findings. Counts only go down. A count of zero, or a workflow missing from the baseline, locks that workflow.
See also
Section titled “See also”Last reviewed: 2026-10-28