colonelpanic8/dotfiles/logical-commits
Claude Code and agent skill by colonelpanic8/dotfiles
- community
Checking that the SKILL.md file is still in the repository.
Use when the user asks to split current git changes into logical commits, clean up commit history, create atomic commits, or stage by hunk. Review the whole worktree, group related changes, and produce ordered commits where each commit is a valid state (builds/tests pass with the project validation command).
Install
npx vskill@latest install colonelpanic8/dotfiles/logical-commitsSKILL.md
Logical Commits
Turn a mixed worktree into a clean sequence of atomic commits.
Workflow
- Inspect the full change set before staging anything.
- Define commit boundaries by behavior or concern, not by file count.
- Order commits so dependencies land first (types/api/schema/helpers before consumers).
- Stage only the exact hunks for one commit.
- Validate that staged commit state is healthy before committing.
- Commit with a precise message.
- Repeat until all intended changes are committed.
1) Inspect First
Run:
If there are staged changes already, inspect both views:
2) Choose Validation Command Early
Select the fastest command that proves the repo is valid for this project. Prefer project-standard commands (for example: just test, npm test, cargo test, go test ./..., nix flake check, targeted build commands).
If no clear command exists:
- Infer the best available command from repo scripts/config.
- Tell the user what command you chose and why.
- Do not claim full validation if coverage is partial.
3) Plan the Commit Stack
Before committing, write a short plan:
- Commit title
- Files and hunks included
- Why this is a coherent unit
- Validation command to run
If changes are intertwined, split by hunk (git add -p). If hunk splitting is not enough, use git add -e or perform a temporary refactor so each commit remains coherent and valid.
4) Stage Exactly One Commit
Preferred staging flow:
Useful corrections:
Never stage unrelated edits just to make the commit pass.
5) Validate Before Commit
Run the chosen validation command with the current staged/working tree state.
If validation fails:
- Fix only what belongs in this logical commit, or
- Unstage/re-split and revise the commit boundary.
Commit only after validation passes.
6) Commit and Verify
Commit:
Then confirm:
Ensure remaining unstaged changes still make sense for later commits.
7) Final Checks
After finishing the stack:
Report:
- The commit sequence created
- Validation command(s) run per commit
- Any residual risks (for example, partial validation only)
Guardrails
- Keep commits atomic and reviewable.
- Prefer hunk staging over broad file staging when a file contains multiple concerns.
- Preserve user changes; do not discard unrelated work.
- Avoid destructive commands unless the user explicitly requests them.
- If a clean logical split is impossible without deeper refactor, explain the blocker and ask for direction.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingdotfiles/agents/skills/logical-commits/SKILL.md
Popularity
- Repository stars
- 220
- Repository forks
- 21
- 7-day trend score
- 27
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/colonelpanic8/dotfiles/logical-commits)