Skip to main content
skill CX Codex

committing-scoped-changes — Claude Code skill

committing-scoped-changes splits an explicit list of files or folders into separate git commits and leaves the rest of the working tree alone.

no license 153 installs ★ 3 updated 2026-10-25

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add b-mendoza/agent-skills --skill committing-scoped-changes -a claude-code
Open on GitHub

What committing-scoped-changes does

You hand the skill a set of paths: exact files, or folders written with a trailing slash. Globs are rejected, spelling has to match the repository, and a rename counts only when both sides are listed.

It runs in two stages. A read-only subagent called commit-boundary-planner reads the state of the repo and writes a plan; you approve that plan in one go; then scoped-commit-executor commits one group at a time with git commit --only -- <paths>. Entries staged from elsewhere stay staged. The report format is checked by scripts/validate-output.sh.

Nothing gets committed until you ask for it in words in the current conversation. An unfinished merge, rebase, cherry-pick or bisect stops the run. Pushing and amending are not part of what it does.

What it helps with

  • commit only the files under src/payments/ without disturbing what is already staged
  • split a pile of edits into separate atomic commits
  • keep unrelated experiments out of a ticket commit
  • verify that no stray file slipped into the commit

Example prompts for the agent

  • Commit only the changes in src/payments/ and tests/payments.test.ts, leave the rest alone.
  • Split my working tree into atomic commits for JNS-6880.
  • Make the commit for src/checkout in Conventional Commits style.

How to install committing-scoped-changes

1Skills CLI (Vercel)

npx skills add b-mendoza/agent-skills --skill committing-scoped-changes -a claude-code

2Manually: put the files in place

~/.claude/skills/committing-scoped-changes/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/committing-scoped-changes/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about committing-scoped-changes

How do I install committing-scoped-changes in Claude Code?

Run npx skills add b-mendoza/agent-skills --skill committing-scoped-changes -a claude-code; the skill lands at ~/.claude/skills/committing-scoped-changes/SKILL.md.

What does committing-scoped-changes do?

It builds atomic git commits from a named set of files or folders and keeps work outside that set out of them.

Can I pass a whole folder or a wildcard?

A folder works when the path ends with /. Wildcards and globs are turned down, and the spelling must match the repository path.

Similar skills

All skills for Git and GitHub →

From the same set12

The whole set →
validate-implementation-plan validate-implementation-plan audits implementation plans for speculative complexity and brittle assumptions. Writes a separate audit file without touching the original plan. CX Codex +1 1,800installs recency-guard recency-guard validates time-sensitive facts in assistant answers, cross-checking software releases, pricing, policy changes, and docs against live sources. CX Codex +1 203installs pr-creator pr-creator turns your current branch into a pull or merge request. You see the exact preview first and approve before anything is created. CX Codex +1 195installs review-pull-request review-pull-request audits a single GitHub pull request across isolated stages, drafts a local Markdown report, and posts comments only when confirmed. CX Codex +1 161installs responding-to-pr-review-comments The responding-to-pr-review-comments skill gathers pull request comments, assesses the feedback, and drafts replies. With confirmation, it can post approved text to supported GitHub comment threads. CX Codex +1 144installs clarifying-assumptions clarifying-assumptions — a skill that runs the conversational layer for plan verification. Asks critical questions upfront or before individual tasks, delegates artifact analysis to subagents. CX Codex +1 129installs refactoring-code refactoring-code is a Claude Code skill that stages simplification refactors: approve the plan before edits, then check that behaviour stays the same. CX Codex +1 124installs diagnosing-root-causes The diagnosing-root-causes skill traces bugs, regressions, and CI/CD failures from supplied evidence, then produces a reviewed root-cause report. Use it when a problem report needs investigation rather than guesswork. CX Codex +1 122installs

Project sponsors

Companies that keep this analytics open