GitHub 深度调研

SkillAI & models

Conduct in-depth research on a GitHub repository (README/structure/key source code/activity) and produce a research report. A pure prompt-based skill that relies on the web_fetch tool.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the GitHub 深度调研 skill

What this skill tells your AI

The instructions your AI receives, as published by yrris/pro-agent in cognition/runtime/skills/github-deep-research/SKILL.md and read by ahel’s review.

调研一个 GitHub 仓库并产出结构化报告。全程用 web_fetch 抓公开页面(无需 token; 配置 COGNITION_GITHUB_TOKEN 后 web_fetch 对 api.github.com / raw.githubusercontent.com 自动携带认证,限流 60→5000 次/小时,抓取写法不变)。

调研流程(按序执行,每步都要真实抓取)

  1. 定位仓库(仅当用户没给出明确的 owner/repo 时):web_search("site:github.com <主题关键词>") → 从结果标题/链接确定 {owner}/{repo} 再进入后续步骤;仓库已明确则跳过本步。
  2. 仓库概览web_fetch("https://api.github.com/repos/{owner}/{repo}") → JSON 含 stars/forks/语言/描述/最近推送时间/开源协议。
  3. READMEweb_fetch("https://raw.githubusercontent.com/{owner}/{repo}/HEAD/README.md") → 项目定位、用法、架构说明。
  4. 目录结构web_fetch("https://api.github.com/repos/{owner}/{repo}/contents/") → 顶层文件/目录 JSON;对关键子目录可再抓 /contents/{path}
  5. 关键源码(按需 2-4 个文件): web_fetch("https://raw.githubusercontent.com/{owner}/{repo}/HEAD/{path}") —— 选入口文件/核心模块/配置(package.json、pyproject.toml、go.mod 判断技术栈)。
  6. 活跃度web_fetch("https://api.github.com/repos/{owner}/{repo}/commits?per_page=5") → 最近提交时间与主题;/issues?state=open&per_page=5 看开放问题。

产出要求

调用 write_report 产出 markdown 报告,结构: 定位与解决的问题 → 技术栈与架构(含目录导览)→ 核心实现要点(引用真实源码路径) → 活跃度与成熟度评估 → 借鉴点/风险点。 所有结论必须来自抓取到的真实内容并注明来源 URL;抓取失败的部分如实说明,不要编造。

详细 URL 模式与备选端点见 references/playbook.md(用 skill_read 查看)。

Signals

GitHub stars
23
Last commit
Jul 2026

ahel review

  • S4info
    community integration — published by yrris, not github

Automated review, not a security audit. Ruleset v1+k2.

Advanced
Catalog kind
skill
Gateway key
github-deep-research-yrris
Source
github.com/yrris/pro-agent