conventional-changelog Committing in a repository that uses commitlint
commitlint validates commit messages against the repository's configured convention (usually Conventional Commits). The configuration is the contract: read it before writing a commit message, and trust the error output when a commit is rejected.
1. Detect whether the repository uses commitlint
Any of these means commitlint is in play:
- A config file in the repository root:
.commitlintrc,.commitlintrc.{json,yaml,yml,js,cjs,mjs,ts,cts,mts}, orcommitlint.config.{js,cjs,mjs,ts,cts,mts} - A
"commitlint"field inpackage.json - A commit-msg hook that runs commitlint (check
.husky/commit-msg,lefthook.yml, or.git/hooks/commit-msg)
2. Read the enforced rules
npx commitlint --print-config json
If this exits non-zero, the configuration itself is broken (for example an
unresolvable extends) — read the error on stderr instead of guessing.
Look at the rules object. Each rule is [severity, applicability, value]:
- severity —
0= disabled,1= warning,2= error (only errors block commits) - applicability —
"always"= the condition must hold,"never"= it must not - value — the rule's parameter (list, number, or string)
The rules that matter most when writing a message:
| Rule | What it controls |
|---|---|
type-enum |
Allowed types (e.g. feat, fix, chore, ...) |
scope-enum |
Allowed scopes (empty list = any scope allowed) |
scope-empty |
[2, "never"] = scope is required; [2, "always"] = scope is forbidden |
subject-case |
With "never": cases the subject must NOT use |
subject-full-stop |
With "never" and ".": subject must not end with a period |
header-max-length |
Maximum length of the first line |
body-leading-blank |
Blank line required between subject and body |
body-max-line-length |
Maximum length of each body line |
A rule that is absent or disabled (severity 0) is simply not enforced. When unsure, use
a conservative baseline that passes the common presets: lower-case type, no trailing
period, header under 72 characters (config-conventional allows up to 100). Note that
running commitlint with no config at all fails with an empty-rules error rather than
falling back to any default convention.
3. Write the message
Format: type(scope): subject — the scope is optional unless scope-empty is [2, "never"].
An example that satisfies @commitlint/config-conventional:
feat(parser): add support for inline issue references
Explain what changed and why. Separate the body from the subject with a blank
line (body-leading-blank) and keep lines within body-max-line-length.
Closes #123
4. Validate before committing
printf '%s' "feat(parser): add support for inline issue references
Explain what changed and why.
Closes #123" | npx commitlint
Exit code 0 with no output means the message passes. Validate the complete message (header and body), not just the first line — body rules are checked too. Doing this before committing is cheaper than a failed commit.
5. Self-correct when the commit-msg hook rejects
Rejection output names each violated rule in brackets:
⧗ --- input ---
Feat: Added new parser.
✖ subject must not be sentence-case, start-case, pascal-case, upper-case [subject-case]
✖ subject may not end with full stop [subject-full-stop]
✖ type must be lower-case [type-case]
✖ type must be one of [build, chore, ci, docs, feat, fix, perf, refactor, revert, style, test] [type-enum]
✖ found 4 problems, 0 warnings
- Fix ONLY the rules named in brackets; keep everything else identical.
- Re-validate via stdin (step 4), then retry the commit.
- NEVER bypass the hook with
git commit --no-verify— fixing the message is always the correct action.