Token导航 LogoToken导航

Agent Skill · 研究检索

github-repo(GitHub repo 搜索)

用于围绕 GitHub 仓库、Issue、Pull Request、分支、提交和代码协作流程提供辅助能力。它适合让 Agent 查询项目状态、整理变更、辅助创建或检查协作事项,并把仓库中的信息转成可执行的下一步。使用时需要区分只读查询和写入操作;涉及创建 PR、修改 Issue、推送分支或访问私有仓库时,应确认 token 权限、目标仓库范围和用户授权。

github-repo 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,权限为只读;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 65,612。

总安装 279周安装 12GitHub Stars 46下载 98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
unknown

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:github-repo(GitHub repo 搜索)
来源仓库:https://github.com/jdevalk/skills
仓库路径:skills/github-repo
安装命令:
npx skills add https://github.com/jdevalk/skills --skill github-repo
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。

skills.shnpx skills
npx skills add https://github.com/jdevalk/skills --skill github-repo

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 GitHub 仓库及相关协作事项的研究检索与信息管理。支持在 Codex、Claude、Cursor 和 Gemini CLI 中调用。通过 GitHub 仓库地址进行技能安装。操作 Issue 或 PR 时需注意 API token 的读写权限配置。

github-repo 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/github-repo

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/jdevalk/skills --skill github-repo;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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

github-repo 怎么安装?

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

github-repo 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。

github-repo 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

github-repo 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

github-repo 用的人多吗?

公开数据是累计安装 279 次,近一周安装 12 次,下载 98 次,GitHub 46 星。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。

github-repo 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

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。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 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
github-repo

GitHub repo 搜索 · 本页

27946只读未明确有提醒

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

GitHub Repository Optimizer

This skill audits a GitHub repository against best practices and generates or improves the files that make a repo look professional, welcoming, and well-maintained. It works with live repos (via gh CLI), local git directories, or files the user provides.

Implementation recipes live in AGENTS.md — read it when you need to generate or improve specific files. This file has the workflow and audit checklist.

Workflow

  1. Gather repo context — Determine what we're working with and collect information
  2. Run the audit — Score every aspect of the repo's public-facing quality
  3. Generate or improve files — Produce the files that are missing or need work. Recipes are in AGENTS.md.
  4. Metadata and readability pass — invoke metadata-check on the repo description and README blurb; readability-check on prose-heavy files.
  5. Verify — Confirm the outputs are correct and complete

Phase 0: Gather Context

Key information to collect: repo name, description, language/framework, README existence and quality, which community health files exist, issue/PR template status, CI/CD status, release history, branch count. Use gh CLI if authenticated, otherwise read local files.


Phase 1: Audit

Score each category out of 10. For each, give 2–4 specific findings that quote the actual state.

README Quality (x/10)

Does the README exist? Is it well-structured with: project name and clear one-liner, badge row (build/license/version), description explaining what/why/who, quick start (install + first use in 3 steps or fewer), code examples, screenshots/demos if visual, configuration docs, contributing link, license mention? Is it under 500 lines with <details> for optional content?

Repository Metadata (x/10)

  • Description set and useful (not blank or "My awesome project")?
  • Topics/tags relevant and complete?
  • Website URL set if applicable?
  • Social preview image uploaded?
  • "Packages", "Environments", "Deployments" sections hidden if unused?

Community Health Files (x/10)

  • CONTRIBUTING.md — exists? Covers dev setup, code style, testing, PR process?
  • SECURITY.md — exists? Directs to private reporting channel, states response time?
  • CODE_OF_CONDUCT.md — exists? Contributor Covenant 2.1 is the standard.
  • LICENSE — exists in root? Appropriate for project type?
  • SUPPORT.md — exists? Separates bugs from support questions?
  • CHANGELOG.md — follows Keep a Changelog format?

For organizations: check for a .github repository with org-wide default health files.

Issue & PR Templates (x/10)

  • YAML-based issue forms (.github/ISSUE_TEMPLATE/*.yml) with at least bug report and feature request? Numbered filenames for ordering?
  • config.yml disabling blank issues and linking external support?
  • PR template with description, related issues (Closes #XXX), type-of-change checkboxes, testing checklist?
  • CODEOWNERS file defining reviewers per directory?

CI/CD & Automation (x/10)

  • GitHub Actions workflows for testing, linting, building? Passing status badges in README?
  • Dependabot or Renovate configured?
  • Code quality (CodeQL, linting actions)?
  • Stale issue management?

Releases & Branch Hygiene (x/10)

  • Tagged releases with changelogs? Semantic versioning?
  • Stale branches cleaned up?
  • .gitattributes with export-ignore?
  • Auto-release (semantic-release or release-please)?

Phase 2: Generate or Improve Files

Based on the audit, generate the files that are missing or need improvement. Always ask before overwriting. Read AGENTS.md for detailed recipes.

File generation order by impact: README.md, LICENSE, CONTRIBUTING.md, SECURITY.md, CODE_OF_CONDUCT.md, issue templates, PR template, CHANGELOG.md, SUPPORT.md, config files (dependabot.yml, CODEOWNERS, FUNDING.yml).

AGENTS.md sections: README generation, Community health files (CONTRIBUTING, SECURITY, CODE_OF_CONDUCT), Issue templates, Organization-level setup (.github repo, org profile, private member profile, domain verification, rulesets).


Phase 2.5: Metadata and readability pass

  1. Metadata pass — metadata-check skill. Run on the repo description (first ~100 chars show in search and profile cards) and the README's opening blurb.
  2. Prose pass — readability-check skill. Run on every prose-heavy file — at minimum the README, plus CONTRIBUTING.md and SECURITY.md if you produced them.

Apply fixes directly. Skip for code-heavy files (issue templates, workflow YAML).


Phase 3: Verify

  • Verify files are in the correct locations.
  • Check that all internal links in the README work.
  • Confirm badge URLs are correctly formatted.
  • If gh is available, remind the user to update the repo's About section (description, topics, website URL, social preview) — these can't be set via files alone.
  • Summarize what was created/changed and what the user should do next.

Output format

Score table (6 categories, x/60), grouped findings quoting actual state, list of files generated/changed, and next steps for the user.

平台分布

Codex

36.17%
按下载量换算35

Claude

30.14%
按下载量换算30

Cursor

18.64%
按下载量换算18

Gemini CLI

9.66%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills