Token导航 LogoToken导航

Agent Skill · 研究检索

open-source-contribution(开源贡献)

open-source-contribution 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。

open-source-contribution 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证为 MIT,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 68,490。

总安装 233周安装 10GitHub Stars 公开资料未说明下载 82核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

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:open-source-contribution(开源贡献)
来源仓库:https://github.com/ronantakizawa/skill-open-source-contribution
仓库路径:skills/open-source-contribution
安装命令:
npx skills add https://github.com/ronantakizawa/skill-open-source-contribution --skill open-source-contribution
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/ronantakizawa/skill-open-source-contribution --skill open-source-contribution

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景或来源线索快速定位候选结果。通过 npx skills add 命令从指定 GitHub 仓库安装并使用。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。

open-source-contribution 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/open-source-contribution

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/ronantakizawa/skill-open-source-contribution -…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

open-source-contribution 怎么安装?

在终端执行 npx skills add https://github.com/ronantakizawa/skill-open-source-contribution --skill ope…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

open-source-contribution 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

open-source-contribution 需要什么权限?

权限级别是接触敏感数据。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。

open-source-contribution 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

open-source-contribution 用的人多吗?

公开数据是累计安装 233 次,近一周安装 10 次,下载 82 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

open-source-contribution 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
open-source-contribution

开源贡献 · 本页

233—接触敏感数据MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Contributing to Open Source

Quick Navigation

TopicDescription
Finding ProjectsResources and labels for finding issues
Understanding CodebasesHow to navigate unfamiliar code
Writing PRsBranch strategy, commits, descriptions
Code ReviewHow to handle feedback
Communicating with MaintainersPrinciples and templates
Maintainer PerspectiveWhat maintainers want
Common MistakesAnti-patterns to avoid
What Counts as MeaningfulHigh-value contributions

Reference Files:

Templates:


Finding Projects

Resources

Labels to Look For

LabelMeaning
good first issueGitHub's official beginner label
first-timers-onlyReserved for first-time contributors
help wantedMaintainers actively seeking help
documentationOften good entry points

Evaluating a Project

Before contributing, verify:

  • Has an OSI-approved open source license
  • Recent commits (within last 3 months)
  • Maintainers respond to issues/PRs
  • Has CONTRIBUTING.md or contribution guidelines
  • Automated tests (CI/CD) exist

Understanding Large Codebases

Step 1: Documentation First

Read in order: README.md → CONTRIBUTING.md → Architecture docs → API docs

Step 2: Build and Run Locally

git clone https://github.com/<owner>/<repo>.git
cd <repo>
# Follow setup instructions, run tests

Step 3: Explore Strategically

Find important files:

git log --pretty=format: --name-only | sort | uniq -c | sort -rg | head -20

Understand entry points: Look for main, index, app, or server files.

Read tests: They document expected behavior and show component usage.

Step 4: Focus on Your Target

Don't try to understand everything:

  1. Identify the module related to your issue
  2. Read that module thoroughly
  3. Trace dependencies one level up and down
  4. Treat unrelated code as "black boxes"

Step 5: Use Git as Documentation

git log --oneline <file>           # File history
git log --grep="<keyword>"         # Find relevant PRs
git blame <file>                   # Who to ask

Writing Effective Pull Requests

Before You Start

  1. Claim the issue: Comment to let maintainers know you're working on it
  2. Ask questions: If requirements are unclear, ask before coding
  3. Check for duplicates: Search existing PRs

Forking Workflow

For projects where you don't have write access:

# 1. Fork on GitHub, then clone your fork
git clone https://github.com/YOUR-USERNAME/<repo>.git
cd <repo>

# 2. Add upstream remote
git remote add upstream https://github.com/ORIGINAL-OWNER/<repo>.git

# 3. Keep fork updated
git fetch upstream
git checkout main
git merge upstream/main

# 4. Create feature branch and work
git checkout -b fix/your-fix
# ... make changes ...
git push origin fix/your-fix

# 5. Open PR from your fork to upstream

Branch Strategy

git checkout -b <type>/<short-description>

# Examples:
git checkout -b fix/null-pointer-exception
git checkout -b feat/add-dark-mode
git checkout -b docs/update-readme

Commit Messages

Follow Conventional Commits:

<type>[scope]: <description>

[optional body]

[optional footer]

Quick reference:

TypeUse For
featNew feature
fixBug fix
docsDocumentation
refactorCode restructure
testTests
choreMaintenance

Example:

fix(auth): resolve token refresh race condition

Fixes #123

For complete guide: See reference/conventional-commits.md

PR Description Template

## Summary
Brief description of what this PR does and why.

## Changes
- Change 1
- Change 2

## Related Issues
Fixes #123

## Testing
- [ ] Existing tests pass
- [ ] Added new tests
- [ ] Manually tested

PR Sizing Best Practices

Research shows smaller PRs get better reviews:

SizeLines ChangedReview Quality
Ideal~50 linesThorough review
Good<200 linesGood feedback
Acceptable<400 linesAdequate review
Too Large400+ linesLikely to miss issues

PR Best Practices

  • One concern per PR: Don't mix features, fixes, and refactors
  • Self-review first: Read your own diff before requesting review
  • Use draft PRs: Open early for feedback on approach
  • Include tests: Maintainers rarely merge untested code
  • Respond promptly: Don't let PRs go stale

What NOT to Include in PRs

Remove before committing:

  • .env, .env.local, credentials, API keys
  • IDE/editor configs (.idea/, .vscode/ unless project-standard)
  • Personal notes, TODOs, planning files
  • Debug code, console.logs, print statements
  • Unrelated formatting changes
  • Large binary files, screenshots (unless required)

Check for secrets before pushing:

# Search for common secret patterns
git diff --cached | grep -iE "(api_key|password|secret|token).*="

GitHub CLI Essentials

# Create PR interactively
gh pr create

# Create PR with title and body
gh pr create --title "Fix: resolve null pointer" --body "Fixes #123"

# Create draft PR
gh pr create --draft

# Check PR status
gh pr status

# View PR in browser
gh pr view --web

Responding to Code Review

Mindset

Code review is collaborative, not adversarial. Reviewers want to help improve the code.

Response Templates

When you agree:

Good catch! Fixed in [commit hash].

When you need clarification:

I want to make sure I understand—are you suggesting [X] because of [Y]?

When you disagree:

I went with [current approach] because:
- [Reason 1]
- [Reason 2]

Open to changing if you think [alternative] better serves the project.

When asked for big changes:

Great suggestion. Would it make sense to address this in a follow-up PR?

Guidelines

  • Address all comments
  • Stay calm—step away if frustrated
  • Re-request review after addressing feedback

Communicating with Maintainers

Principles

  • Keep communication public (others benefit)
  • Be concise (maintainers have limited time)
  • Do homework first (search existing issues)
  • Be patient (follow up after one week, politely)

Bug Report Template

## Description
Clear description of the bug.

## Steps to Reproduce
1. Step 1
2. Step 2
3. See error

## Expected vs Actual Behavior
Expected: X
Actual: Y

## Environment
- OS: [e.g., macOS 14.0]
- Version: [e.g., v2.1.0]

Feature Request Template

## Summary
What you're proposing.

## Problem
What problem this solves.

## Proposed Solution
How you envision it working.

## Alternatives Considered
Other approaches you considered.

Understanding Maintainer Perspective

What Maintainers Deal With

  • Overwhelm: Popular projects receive hundreds of issues/PRs
  • Volunteer work: Most maintainers aren't paid
  • Burnout: Endless notifications and demanding users
  • Quality gates: Must protect codebase from bugs and technical debt

What Maintainers Want

  1. Contributors who read the docs first
  2. Well-tested code
  3. Clear communication about what and why
  4. Patience (days or weeks to respond is normal)
  5. Follow-through (don't abandon PRs mid-review)

Quotes from Experienced Maintainers

"The best good first issue is the one you created yourself. Try going through the product, and in the process of testing and understanding it, you'll find your good first issue."
"All the projects I've contributed to are things I've used in some way. I never saw the point of just 'showing up' to a project."

Common Mistakes to Avoid

Before Starting

  1. Skipping CONTRIBUTING.md: Always read contribution guidelines first
  2. Not checking existing work: Search PRs/issues for duplicates
  3. Working on assigned issues: Check if someone is already on it
  4. Building unsolicited features: Propose in an issue first, wait for approval
  5. Not understanding the project: Use it before contributing to it

During Development

  1. Working on main branch: Always use feature branches
  2. Giant PRs: Break into smaller, focused PRs (<200 lines ideal)
  3. No tests: Maintainers rarely merge untested code
  4. Including secrets: Check for API keys, passwords, tokens
  5. Committing debug code: Remove console.logs and print statements

When Submitting

  1. Vague PR titles: Be specific (not "Fixed bug" but "Fix null pointer in auth handler")
  2. Ignoring templates: Use provided issue/PR templates
  3. Ignoring CI failures: Fix all failing checks before requesting review
  4. No description: Explain what, why, and how to test

After Submitting

  1. Going silent: Respond to feedback within 48 hours
  2. Arguing in reviews: Stay collaborative, assume good intent

Low-Value Contributions

  • Adding your name to README files
  • Trivial changes (typo fixes in comments nobody reads)
  • "Drive-by" contributions with no intention to follow through
  • Opening issues that are already documented

What Counts as Meaningful

High-Value

  • Bug fixes with tests
  • Documentation that helps new users
  • Test coverage for untested code paths
  • Performance improvements with benchmarks
  • Security fixes (reported responsibly)
  • Triaging issues: Reproducing bugs, closing duplicates

Best Strategy

  1. Use the project first: Become a real user before contributing
  2. Solve your own problems: Fix bugs you encounter
  3. Think long-term: Build relationships, not just contribution counts
  4. Quality over quantity: One thoughtful PR beats ten trivial ones

External References

Official Guides

Finding Projects

Community Insights

平台分布

Claude Code

26.23%
按下载量换算22

OpenCode

22.06%
按下载量换算18

windsurf

18.06%
按下载量换算15

trae

13.21%
按下载量换算11

Codex

8.89%
按下载量换算7

Antigravity

3.77%
按下载量换算3

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

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。

继续浏览同类 Skills