jesseleite/dotfiles/commit

Claude Code and agent skill by jesseleite/dotfiles

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Plan a sequenced set of commits matching repo message style, then land them one at a time with staged diffs previewed in Hunk to the right of the agent. Use when the user wants to break work into commits, land a series of commits, or runs /commit.

Install

npx vskill@latest install jesseleite/dotfiles/commit

SKILL.md

When asked to break work into commits or run /commit, do the following in order:

  1. Plan commits (do not stage yet):
  • Inspect the worktree and recent commits on this branch/repo
  • Propose a sensible sequence of commits
  • Ideally not too small or too big
  • Find the right balance between atomic and related changes, by functionality and intent
  • Don't be concerned about the number of commits if there are a lot of unstaged changes
  • Match commit atomicity and message style already used on this branch
  • Only sample the user's own commits for style (eg. git log --author=...)
  • Do not glean style from bot-generated commits or merge commits
  • No ticket numbers in commit subjects unless specifically referencing a different ticket
  • If helpful, include relevant context, reasoning, and decision making in the commit body
  • Wait for the user to agree on the high-level plan before staging anything
  • Grill the user if sequencing or scope is genuinely ambiguous
  1. For each commit, one at a time:
  • Git stage only that commit's changes
  • Open (or reuse, if one exists) a Herdr pane to the right of the current agent pane
  • Show the staged diff there with Hunk (eg. hunk diff --staged --watch --sidebar)
  • Tell the user the commit number, proposed message, and that Hunk is ready on the right
  • Always refresh the Hunk diff and clear review notes at this point
  • Do not steal focus from the agent pane while they may still be typing
  • Stop. Wait for approval, message edits, or Hunk notes
  • If they leave Hunk notes:
  • Address any changes that need to be made
  • Grill the user if further clarification is needed
  • Feel free to reply to specific user notes as agent notes in Hunk
  • Re-stage relevant changes and wait again
  • With hunk --watch flag, you shouldn't need to refresh hunk for user to see changes
  • Commit only when the user is happy and explicitly says to
  • Never git push unless explicitly asked (and NEVER push to main/master/default branch)
  1. After each successful commit, immediately stage and preview the next one the same way until the plan is done (or the user pauses).
  1. When finished:
  • Summarize what landed
  • Leave the Hunk pane alone unless they want it closed

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

Popularity

Repository stars
121
Repository forks
12
7-day trend score
24

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
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/jesseleite/dotfiles/commit/badge)](https://verified-skill.com/skills/jesseleite/dotfiles/commit)

Something wrong with this skill or its page? Tell us.