Token导航 LogoToken导航TokenDH.com
研究检索敏感数据github未标认证来源可访问clear审计未展示

commit提交

Agent Skill

commit 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

343

周安装

14

GitHub Stars

公开资料未说明

下载量

111
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:commit(提交)
来源仓库:https://github.com/arcblock/agent-skills
仓库路径:skills/commit
安装命令:
npx skills add arcblock/agent-skills --skill "commit"
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

AgentSkills.tonpx skills
npx skills add arcblock/agent-skills --skill "commit"

简介

commit 用于发现和安装各类 AI 代理的技能插件。

  • 支持 Codex、Claude、Cursor 和 Gemini CLI 等宿主环境扩展能力。
  • 通过命令行快速添加新技能,简化工具集成流程。commit 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 需具备 GitHub 仓库读取权限和网络连通性。
  • 安装后建议查看原始 README 以了解具体功能与使用边界。

SKILL.md

Commit Skill

Generate standardized git commit messages following the project's Conventional Commits specification.

Commit Message Format

<type>(<scope>): <subject>

<body>

<footer>

Type (Required)

Based on project history analysis, use the following types:

TypeDescriptionExample
featNew featurefeat: add user authentication
fixBug fixfix: resolve login timeout issue
docsDocumentation onlydocs: update API documentation
styleCode style (formatting, semicolons, etc.)style: fix indentation
refactorCode refactoring (no feature change)refactor: extract validation logic
perfPerformance improvementperf: optimize database queries
testAdd or modify teststest: add unit tests for auth module
choreBuild process or auxiliary toolschore: update dependencies
ciCI/CD configurationci: add GitHub Actions workflow
buildBuild system changesbuild: upgrade webpack to v5

Scope (Optional)

Scope indicates the affected module or area, enclosed in parentheses:

feat(devflow): add diff-review-doc skill
fix(auth): resolve token refresh issue

Common scopes in this project:

  • Plugin names: devflow, blocklet, arcblock-context
  • Skill names: blocklet-dev-setup, blocklet-pr
  • Module names: specific functional modules

Subject (Required)

  • Use imperative mood: "add" not "added" or "adds"
  • Lowercase first letter
  • No period at the end
  • Keep under 50 characters
  • Describe what the commit does

Good examples:

  • feat: add blocklet-branch skill for branch management
  • fix: resolve gh command order issue
  • docs: update installation guide

Bad examples:

  • feat: Added new feature. (wrong tense, has period)
  • Fix bug (missing type prefix)
  • feat: This commit adds a new feature that... (too verbose)

Body (Optional)

  • Wrap at 72 characters
  • Explain "what" and "why" vs "how"
  • Separate from subject with blank line

Footer (Optional)

  • Reference issues: Closes #123
  • Breaking changes: BREAKING CHANGE: description

Workflow

When user requests to create a commit:

Step 1: Check Current Status

git status
git diff --staged
git diff

Step 2: Verify All Changes Are Staged

IMPORTANT: If there are unstaged changes, use AskUserQuestion to confirm how to proceed.

Check for unstaged changes:

  • Modified files not staged (Changes not staged for commit)
  • Untracked files (Untracked files)

If unstaged changes exist:

  1. List all unstaged/untracked files
  2. Use AskUserQuestion tool to ask user how to proceed:
{
  "questions": [{
    "question": "There are unstaged changes. How would you like to proceed?",
    "header": "Unstaged",
    "options": [
      {"label": "Stage all", "description": "Run `git add .` to stage all changes"},
      {"label": "Abort", "description": "Cancel the commit and handle changes manually"}
    ],
    "multiSelect": false
  }]
}
  1. If user selects "Stage all": run git add. then continue to Step 3
  2. If user selects "Abort": stop the workflow and inform user to stage changes manually
  3. DO NOT proceed until all intended changes are staged

Step 3: Analyze Staged Changes

Only proceed if all changes are properly staged:

  • Identify the type of change (feature, fix, docs, etc.)
  • Determine the scope if applicable
  • Summarize the change concisely

Step 4: Generate and Confirm Commit Message

Generate the commit message following the format, then use AskUserQuestion tool to confirm:

Proposed commit message:

<type>(<scope>): <subject>

<body if needed>

Co-Authored-By: Claude <noreply@anthropic.com>

Use AskUserQuestion tool with options:

  • Approve - Proceed to commit
  • Edit message - User wants to modify the message
  • Cancel - Abort the commit

Example AskUserQuestion usage:

{
  "questions": [{
    "question": "Proceed with this commit message?",
    "header": "Commit",
    "options": [
      {"label": "Approve", "description": "Create the commit with this message"},
      {"label": "Edit message", "description": "Modify the commit message"},
      {"label": "Cancel", "description": "Abort the commit"}
    ],
    "multiSelect": false
  }]
}

Step 5: Execute Commit (Only After User Confirmation)

git commit -m "$(cat <<'EOF'
<type>(<scope>): <subject>

<body if needed>

Co-Authored-By: Claude <noreply@anthropic.com>
EOF
)"

Step 6: Verify Commit

git log -1 --oneline

Examples from Project History

feat: parse blocklet urls from dev output instead of constructing from did
feat: add blocklet-branch skill for branch management
feat: enhance blocklet-server-dev-setup skill with GitHub CLI authentication guidance
feat(devflow): add diff-review-doc skill (#1)
fix: gh move top
docs: prefer git URL format for marketplace installation
chore: initial commit
chore: cleanup hello world plugin

Rules

  1. Always use English for commit messages
  2. Always use lowercase type prefix
  3. Always use imperative mood in subject
  4. Never end subject with period
  5. Add Co-Authored-By footer when AI generates the commit
  6. Keep subject under 50 characters when possible
  7. Reference PR/Issue numbers when applicable: (#123)

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenCode

27.63%
按下载量换算31

Codex

20.86%
按下载量换算23

Claude Code

17.18%
按下载量换算19

Antigravity

12.79%
按下载量换算14

Gemini CLI

8.1%
按下载量换算9

windsurf

3.35%
按下载量换算4

安全审计

暂无安全审计结果可展示。

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills