write-commit-messages
Skill7 starsChanged 15 days agoMITView on GitHub
Use when committing work or writing a commit message — produces a subject and body any developer from junior to CTO can read and act on.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- VelimirMueller/claude-skillsword for word, first seen Oct 11, 2026
Other versions1 project uses a version of this.
Word for word · 1
- VelimirMueller/claude-skills★ 7100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository156 other files it gives its agents.
- skills/swift/swiftui-design-principles/AGENTS.mdAGENTS.md
- skills/swift/ios-code-audit/CLAUDE.mdCLAUDE.md
- skills/swift/swift-focusengine-pro/llms.txtllms.txt
- build-llm-seamSkill
- build-mcp-serverSkill
- build-rag-pipelineSkill
- secure-llm-featuresSkill
- set-up-llm-evalsSkill
- build-nextjs-backendSkill
- build-supabase-edge-functionSkill
- configure-backend-testsSkill
- design-http-apiSkill
- harden-backendSkill
- scaffold-fastapi-serviceSkill
- scaffold-go-serviceSkill
- scaffold-hono-serviceSkill
- secure-supabase-rlsSkill
- set-up-backend-authSkill
- set-up-background-jobsSkill
- set-up-databaseSkill
- set-up-nextjs-supabase-authSkill
- set-up-observabilitySkill
- set-up-supabaseSkill
- build-cliSkill
- release-cliSkill
- set-up-dev-toolchainSkill
- audit-securitySkill
- audit-toolchainSkill
- extend-skillsetSkill
- promote-releaseSkill
- set-up-stack-profileSkill
- write-pull-requestsSkill
- analyze-complexitySkill
- apply-algorithmsSkill
- apply-design-patternsSkill
- apply-design-principlesSkill
- apply-solidSkill
- choose-data-structureSkill
- debug-systematicallySkill
- enforce-dependency-ruleSkill
- refactor-safelySkill
- slice-and-estimateSkill
- write-clean-testsSkill
- write-readable-codeSkill
- clean-frontend-scaffoldingSkill
- configure-accessibilitySkill
- configure-analyticsSkill
- configure-ciSkill
- configure-error-trackingSkill
- configure-lintingSkill
- configure-test-stackSkill
- configure-typescriptSkill
- create-moduleSkill
- optimize-performanceSkill
- scaffold-frontend-projectSkill
- set-up-authSkill
- set-up-design-systemSkill
- set-up-document-headSkill
- set-up-error-boundariesSkill
- set-up-feature-flagsSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 0 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
Read the actual diff: the body will be derived from it, never from memory of what was intended.
- Very specific: line 45
Test: "applying this commit will <subject>" reads as a sentence — the description after any type(scope): prefix is what must parse …
- Very specific: line 65
Every bullet concrete — a path, a behavior, a named mechanism, a number — and project jargon glossed on first use.
- Very specific: line 67
Keep the trailers the house convention carries (Co-Authored-By:, Signed-off-by:) after the three groups — they are part of what the step-1 …
- Very specific: line 72
Run the three tests from ../_shared/audience.md: a junior can follow the message without tribal knowledge, a senior can verify each bullet …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Write Commit Messages
The diff already records what changed; the message carries what the diff cannot — the motivation and the approach. Subject in the repository's own convention; body in up to three labeled groups — Why, What, How — every line derived from the diff. Scale to the change: a one-line fix gets a subject and, at most, a Why line.
1. Audit — the diff and the house convention
git diff --staged # or: git show <commit>, for an existing message
git log --oneline -15Read the actual diff: the body will be derived from it, never from memory of what was
intended. From the log, detect the subject convention — conventional commits
(feat(scope): …), gitmoji, or plain imperative — and match it; default to Conventional
Commits 1.0.0 (type(scope)!: description, ! or a BREAKING CHANGE: footer for breaking
changes) only when no convention is detectable. Change nothing yet.
2. Decide
- The diff mixes concerns (a feature plus an unrelated fix) → propose the split before writing anything: one concern per commit, one message each.
- A commit exists and only its message is wrong → amend rather than stack a fixup:
git commit --amend. Safe only while the commit is unpushed and unshared — rewriting published history breaks everyone who pulled it. Already pushed → never amend; when the message is still wanted (a squash target, the spine of a PR description), draft it without touching history. - The change is trivial (typo, rename, dependency bump the subject fully explains) →
subject only; skip step 4. Ceremony on a one-liner is noise in
git log. - The message already conforms (subject in convention, the three groups present) → report "already in place" and exit. A second run writes nothing.
3. Draft the subject
In the detected convention, imperative mood (Git's convention; the Conventional Commits
spec itself is silent on it), aim ≤72 characters (a display limit, not a spec rule), the
one concern named.
Test: "applying this commit will <subject>" reads as a sentence — the description after
any type(scope): prefix is what must parse (./commit-patterns.md for the discipline).
4. Draft the body — Why, What, How
Up to three labeled bullet groups, each grounded in the step-1 diff. Why: is
mandatory; omit What: when the diff and subject already say it, and How: when the
approach is the obvious one:
Why:
- the problem or motivation — the one thing the diff cannot show
What:
- the change at a glance: surface, files, behavior
How:
- the approach taken and its trade-offsEvery bullet concrete — a path, a behavior, a named mechanism, a number — and project
jargon glossed on first use. What separates concrete from filler: ./commit-patterns.md.
Keep the trailers the house convention carries (Co-Authored-By:, Signed-off-by:)
after the three groups — they are part of what the step-1 audit detects.
5. Self-check — the audience contract
Run the three tests from ../_shared/audience.md: a junior can follow the message
without tribal knowledge, a senior can verify each bullet against the diff, a CTO gets
outcome and risk from the labels and first lines alone. Any unexplained jargon or
unverifiable claim → revise before showing it.
6. Verify
Show the final message for approval before committing; after committing, confirm:
git log -1 --format=%BThe printed message matches what was approved: subject in the repo's convention, the three labeled groups present.
References
- ./commit-patterns.md — the rules with rationale, the unguided baseline as anti-example, and a worked rewrite of a real commit.
- ../_shared/audience.md — the junior-to-CTO contract step 5 checks against.
- ../write-pull-requests/SKILL.md — the same Why/What/How spine at PR scale; on squash-merge, the PR description becomes this commit's message.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

