Token导航 LogoToken导航

Agent Skill · 研究检索

gh-work-report(GH 工作报告)

gh-work-report 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。

gh-work-report 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 52,782。

总安装 546周安装 23GitHub Stars 55下载 191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:gh-work-report(GH 工作报告)
来源仓库:https://github.com/rysweet/amplihack
仓库路径:skills/gh-work-report
安装命令:
npx skills add https://github.com/rysweet/amplihack --skill gh-work-report
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/rysweet/amplihack --skill gh-work-report

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。适合根据关键词、任务场景或来源线索快速定位候选结果。安装命令:npx skills add https://github.com/rysweet/amplihack --skill gh-work-report 建议确认权限范围、维护状态及是否会触发联网、命令执行或文件读写。

可结合原始 README 继续核验具体用法。

来源仓库内路径:skills/gh-work-report

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/rysweet/amplihack --skill gh-work-report;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

gh-work-report 怎么安装?

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

gh-work-report 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。

gh-work-report 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。

gh-work-report 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。

gh-work-report 用的人多吗?

公开数据是累计安装 546 次,近一周安装 23 次,下载 191 次,GitHub 55 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。

gh-work-report 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。

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。
  • 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。

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
gh-work-report

GH 工作报告 · 本页

54655接触敏感数据未明确有提醒

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 Work Report Skill

Generates rich GitHub activity reports across all authenticated gh accounts.

Invocation

The user triggers this skill with phrases like:

  • /gh-work-report (default: last 7 days)
  • /gh-work-report show me the last 30 days
  • generate a github work report for the last 90 days
  • what did I work on this week

Parse the time period

Extract the number of days from the user's message. Valid values: 1, 5, 7, 30, 90. Default to 7 if not specified or if an invalid value is given.

Workflow

Follow these steps in order. Check off each step as you complete it.

Step 1: Detect accounts

gh auth status 2>&1

Parse the output to identify all authenticated accounts (e.g., rysweet on github.com, rysweet_microsoft on github.com). Store the list of account values and note which is currently active.

Step 2: Gather data per account

For each account, switch to it and collect data:

gh auth switch --user <ACCOUNT>

Then gather:

2a. Repositories with recent activity
# Get repos the user pushed to in the time window
gh api graphql --paginate -f query='
query($cursor: String) {
  viewer {
    repositories(first: 100, after: $cursor, orderBy: {field: PUSHED_AT, direction: DESC}) {
      pageInfo { hasNextPage endCursor }
      nodes {
        nameWithOwner
        url
        description
        pushedAt
        homepageUrl
        isPrivate
      }
    }
  }
}'

Filter to repos with pushedAt within the time window.

2b. Pull requests
gh search prs --author=@me --created=">YYYY-MM-DD" --limit 200 --json number,title,repository,state,createdAt,url,mergedAt

Also gather PRs merged (not just created) in the window:

gh search prs --author=@me --merged=">YYYY-MM-DD" --limit 200 --json number,title,repository,state,createdAt,url,mergedAt

Deduplicate by URL.

2c. Issues
gh search issues --author=@me --created=">YYYY-MM-DD" --limit 100 --json number,title,repository,state,createdAt,url
2d. Releases
# For each active repo, check for releases
gh api repos/{owner}/{repo}/releases --jq '.[].tag_name' | head -5

Step 3: Combine and deduplicate

Merge data from all accounts. Deduplicate repos by nameWithOwner and PRs by URL. Tag each item with the account that produced it.

Step 4: Analyze and synthesize

This is where you add value beyond raw data:

  1. Identify themes: Group repos/PRs by topic (e.g., "infrastructure", "security", "new features"). Use repo descriptions, PR titles, and any patterns you observe.
  2. Highlight big wins: PRs with significant impact — large features merged, important bug fixes, new repos created.
  3. Extract usage examples: For notable features, write a short "here's how to use this" snippet based on PR titles, descriptions, and repo READMEs.
  4. Spot new work: Repos with first-ever commits in the time window.

Step 5: Generate the report

Use the template structure from reference.md. The report must include:

  • Executive summary (3-5 sentences)
  • Activity overview with mermaid charts
  • Per-project sections with PR tables
  • Themes and big wins
  • Usage examples for notable features
  • Appendix with raw data links

Save the report as a markdown file named gh-work-report-YYYY-MM-DD-to-YYYY-MM-DD.md.

Step 6: Restore original account

Switch back to the account that was active before the report started:

gh auth switch --user <ORIGINAL_ACCOUNT>

Step 7: Offer automation infrastructure

After generating the report, ask the user:

Would you like me to create a private GitHub repo with automated weekly/monthly reports and a GitHub Pages site to browse them?

If yes, follow the infrastructure setup in reference.md § Infrastructure Setup.

Key Rules

  • Never hardcode usernames — always detect from gh auth status
  • Never use fallbacks — no silent defaults, no 2>/dev/null, no hardcoded values. Errors must fail loud with descriptive messages.
  • All charts must be code-generated — every number in every chart and table must come from actual API data. Never fabricate or estimate chart data.
  • 4 query filters for complete PR coverage: created:>DATE (new), is:open (all WIP), merged:>DATE (merged during window), closed:>DATE (closed during window). Deduplicate by URL.
  • Clamp chart timelines — Gantt and timeline charts must be scoped to the report window. Clamp start dates to max(pr_date, window_start).
  • Handle private repos gracefully — note them but don't expose sensitive details unless the report itself is private
  • Date math: Use date -d "$DAYS days ago" +%Y-%m-%d (Linux) for the start date
  • Rate limiting: If gh api returns 403, wait and retry. Use --paginate for large result sets.
  • Empty results are fine — if an account has no activity, say so briefly and move on

Authentication & Automation

  • Local (multi-account): Use ./run.sh [days] which leverages gh auth switch across all locally authenticated accounts. This is the recommended approach for users with multiple accounts (e.g., public + EMU).
  • GitHub Actions (single-account): Use the workflow templates with a PAT secret. A PAT is scoped to one identity and cannot cross accounts.
  • Two-PAT approach: For Actions across two accounts, use separate PAT secrets (ACCOUNT1_PAT, ACCOUNT2_PAT) with explicit --header "authorization: token $PAT" per API call.
  • GitHub Pages is static only — report generation must happen elsewhere (local script, Actions, gh-aw). Pages just serves the output.

Reference Files

  • reference.md — Report template, mermaid chart patterns, infrastructure setup guide
  • templates/weekly-report.yml — GitHub Actions workflow for automated reports
  • templates/pages-index.html — GitHub Pages aggregation site template

平台分布

Codex

34.71%
按下载量换算66

Claude

29.21%
按下载量换算56

Cursor

17.82%
按下载量换算34

Gemini CLI

9.44%
按下载量换算18

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

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。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