← 返回 Skills 市场
GitHub项目分析助手
作者
flyingants
· GitHub ↗
· v1.0.0
· MIT-0
153
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install github-explorer
功能描述
Deep-dive analysis of GitHub projects. Use when the user mentions a GitHub repo/project name and wants to understand it — triggered by phrases like "帮我看看这个项目...
使用说明 (SKILL.md)
\r \r
GitHub Explorer — 项目深度分析\r
\r
Philosophy: README 只是门面,真正的价值藏在 Issues、Commits 和社区讨论里。\r \r
Workflow\r
\r
[项目名] → [1. 定位 Repo] → [2. 多源采集] → [3. 分析研判] → [4. 结构化输出]\r
```\r
\r
### Phase 1: 定位 Repo\r
\r
- 用 `web_search` 搜索 `site:github.com \x3Cproject_name>` 确认完整 org/repo\r
- 用 `search-layer`(Deep 模式 + 意图感知)补充获取社区链接和非 GitHub 资源:\r
```bash\r
python3 skills/search-layer/scripts/search.py \\r
--queries "\x3Cproject_name> review" "\x3Cproject_name> 评测 使用体验" \\r
--mode deep --intent exploratory --num 5\r
```\r
- 用 `web_fetch` 抓取 repo 主页获取基础信息(README、Stars、Forks、License、最近更新)\r
\r
### Phase 2: 多源采集(并行)\r
\r
**⚠️ GitHub 页面抓取规则(强制)**:GitHub repo 页面是 SPA(客户端渲染),`web_fetch` 只能拿到导航栏壳子,**禁止用 web_fetch 抓 github.com 的 repo 页面**。一律使用 GitHub API:\r
- README: `curl -s -H "Authorization: token {PAT}" -H "Accept: application/vnd.github.v3.raw" "https://api.github.com/repos/{owner}/{repo}/readme"`\r
- Repo 元数据: `curl -s -H "Authorization: token {PAT}" "https://api.github.com/repos/{owner}/{repo}"`\r
- Issues: `curl -s -H "Authorization: token {PAT}" "https://api.github.com/repos/{owner}/{repo}/issues?state=all&sort=comments&per_page=10"`\r
- Commits: `curl -s -H "Authorization: token {PAT}" "https://api.github.com/repos/{owner}/{repo}/commits?per_page=10"`\r
- File tree: `curl -s -H "Authorization: token {PAT}" "https://api.github.com/repos/{owner}/{repo}/git/trees/{branch}?recursive=1"`\r
\r
PAT 见 TOOLS.md。\r
\r
以下来源**按需检查**,有则采集,无则跳过:\r
\r
| 来源 | URL 模式 | 采集内容 | 建议工具 |\r
|---|---|---|---|\r
| GitHub Repo | `github.com/{org}/{repo}` | README、About、Contributors | `web_fetch` |\r
| GitHub Issues | `github.com/{org}/{repo}/issues?q=sort:comments` | Top 3-5 高质量 Issue | `browser` |\r
| 中文社区 | 微信/知乎/小红书 | 深度评测、使用经验 | `content-extract` |\r
| 技术博客 | Medium/Dev.to | 技术架构分析 | `web_fetch` / `content-extract` |\r
| 讨论区 | V2EX/Reddit | 用户反馈、槽点 | `search-layer`(Deep 模式) |\r
\r
#### search-layer 调用规范\r
\r
search-layer v2 支持意图感知评分。github-explorer 场景下的推荐用法:\r
\r
| 场景 | 命令 | 说明 |\r
|------|------|------|\r
| **项目调研(默认)** | `python3 skills/search-layer/scripts/search.py --queries "\x3Cproject> review" "\x3Cproject> 评测" --mode deep --intent exploratory --num 5` | 多查询并行,按权威性排序 |\r
| **最新动态** | `python3 skills/search-layer/scripts/search.py "\x3Cproject> latest release" --mode deep --intent status --freshness pw --num 5` | 优先新鲜度,过滤一周内 |\r
| **竞品对比** | `python3 skills/search-layer/scripts/search.py --queries "\x3Cproject> vs \x3Ccompetitor>" "\x3Cproject> alternatives" --mode deep --intent comparison --num 5` | 对比意图,关键词+权威双权重 |\r
| **快速查链接** | `python3 skills/search-layer/scripts/search.py "\x3Cproject> official docs" --mode fast --intent resource --num 3` | 精确匹配,最快 |\r
| **社区讨论** | `python3 skills/search-layer/scripts/search.py "\x3Cproject> discussion experience" --mode deep --intent exploratory --domain-boost reddit.com,news.ycombinator.com --num 5` | 加权社区站点 |\r
\r
**意图类型速查**:`factual`(事实) / `status`(动态) / `comparison`(对比) / `tutorial`(教程) / `exploratory`(探索) / `news`(新闻) / `resource`(资源定位)\r
\r
> 不带 `--intent` 时行为与 v1 完全一致(无评分,按原始顺序输出)。\r
\r
降级规则:Exa/Tavily 任一 429/5xx → 继续用剩余源;脚本整体失败 → 退回 `web_search` 单源。\r
\r
---\r
\r
### 抓取降级与增强协议 (Extraction Upgrade)\r
\r
当遇到以下情况时,**必须**从 `web_fetch` 升级为 `content-extract`:\r
1. **域名限制**: `mp.weixin.qq.com`, `zhihu.com`, `xiaohongshu.com`。\r
2. **结构复杂**: 页面包含大量公式 (LaTeX)、复杂表格、或 `web_fetch` 返回的 Markdown 极其凌乱。\r
3. **内容缺失**: `web_fetch` 因反爬返回空内容或 Challenge 页面。\r
\r
调用方式:\r
```bash\r
python3 skills/content-extract/scripts/content_extract.py --url \x3CURL>\r
```\r
\r
content-extract 内部会:\r
- 先检查域名白名单(微信/知乎等),命中则直接走 MinerU\r
- 否则先用 `web_fetch` 探针,失败再 fallback 到 MinerU-HTML\r
- 返回统一 JSON 合同(含 `ok`, `markdown`, `sources` 等字段)\r
\r
### Phase 3: 分析研判\r
\r
基于采集数据进行判断:\r
\r
- **项目阶段**: 早期实验 / 快速成长 / 成熟稳定 / 维护模式 / 停滞(基于 commit 频率和内容)\r
- **精选 Issue 标准**: 评论数多、maintainer 参与、暴露架构问题、或包含有价值的技术讨论\r
- **竞品识别**: 从 README 的 "Comparison"/"Alternatives" 章节、Issues 讨论、以及 web 搜索中提取\r
\r
### Phase 4: 结构化输出\r
\r
严格按以下模板输出,**每个模块都必须有实质内容或明确标注"未找到"**。\r
\r
#### 排版规则(强制)\r
\r
1. **标题必须链接到 GitHub 仓库**(格式:`# [Project Name](https://github.com/org/repo)`,确保可点击跳转)\r
2. **标题前后都统一空行**(上一板块结尾 → 空行 → 标题 → 空行 → 内容,确保视觉分隔清晰)\r
3. **Telegram 空行修复(强制)**:Telegram 会吞掉列表项(`-` 开头)后面的空行。解决方案:在列表末尾与下一个标题之间,插入一行盲文空格 `⠀`(U+2800),格式如下:\r
```\r
- 列表最后一项\r
\r
⠀\r
**下一个标题**\r
```\r
这确保在 Telegram 渲染时标题前的空行不被吞掉。\r
2. **所有标题加粗**(emoji + 粗体文字)\r
3. **竞品对比必须附链接**(GitHub / 官网 / 文档,至少一个)\r
4. **社区声量必须具体**:引用具体的帖子/推文/讨论内容摘要,附原始链接。不要写"评价很高"、"热度很高"这种概括性描述,要写"某某说了什么"或"某帖讨论了什么具体问题"\r
5. **信息溯源原则**:所有引用的外部信息都应附上原始链接,让读者能追溯到源头\r
\r
```markdown\r
# [{Project Name}]({GitHub Repo URL})\r
\r
**🎯 一句话定位**\r
\r
{是什么、解决什么问题}\r
\r
**⚙️ 核心机制**\r
\r
{技术原理/架构,用人话讲清楚,不是复制 README。包含关键技术栈。}\r
\r
**📊 项目健康度**\r
\r
- **Stars**: {数量} | **Forks**: {数量} | **License**: {类型}\r
- **团队/作者**: {背景}\r
- **Commit 趋势**: {最近活跃度 + 项目阶段判断}\r
- **最近动态**: {最近几条重要 commit 概述}\r
\r
**🔥 精选 Issue**\r
\r
{Top 3-5 高质量 Issue,每条包含标题、链接、核心讨论点。如无高质量 Issue 则注明。}\r
\r
**✅ 适用场景**\r
\r
{什么时候该用,解决什么具体问题}\r
\r
**⚠️ 局限**\r
\r
{什么时候别碰,已知问题}\r
\r
**🆚 竞品对比**\r
\r
{同赛道项目对比,差异点。每个竞品必须附 GitHub 或官网链接,格式示例:}\r
- **vs [GraphRAG](https://github.com/microsoft/graphrag)** — 差异描述\r
- **vs [RAGFlow](https://github.com/infiniflow/ragflow)** — 差异描述\r
\r
**🌐 知识图谱**\r
\r
- **DeepWiki**: {链接或"未收录"}\r
- **Zread.ai**: {链接或"未收录"}\r
\r
**🎬 Demo**\r
\r
{在线体验链接,或"无"}\r
\r
**📄 关联论文**\r
\r
{arXiv 链接,或"无"}\r
\r
**📰 社区声量**\r
\r
**X/Twitter**\r
\r
{具体引用推文内容摘要 + 链接,格式示例:}\r
- [@某用户](链接): "具体说了什么..."\r
- [某讨论串](链接): 讨论了什么具体问题...\r
{如未找到则注明"未找到相关讨论"}\r
\r
**中文社区**\r
\r
{具体引用帖子标题/内容摘要 + 链接,格式示例:}\r
- [知乎: 帖子标题](链接) — 讨论了什么\r
- [V2EX: 帖子标题](链接) — 讨论了什么\r
{如未找到则注明"未找到相关讨论"}\r
\r
**💬 我的判断**\r
\r
{主观评价:值不值得投入时间,适合什么水平的人,建议怎么用}\r
```\r
\r
## Execution Notes\r
\r
- 优先使用 `web_search` + `web_fetch`,browser 作为备选\r
- **搜索增强**:项目调研类任务默认使用 `search-layer` v2 Deep 模式 + `--intent exploratory`(Brave + Exa + Tavily 三源并行去重 + 意图感知评分),单源失败不阻塞主流程\r
- **抓取降级(强制)**:当 `web_fetch` 失败/403/反爬页/正文过短,或来源域名属于高风险站点(如微信/知乎/小红书)时:改用 `content-extract`(其内部会 fallback 到 MinerU-HTML),拿到更干净的 Markdown + 可追溯 sources\r
- 并行采集不同来源以提高效率\r
- 所有链接必须真实可访问,不要编造 URL\r
- 中文输出,技术术语保留英文\r
\r
## ⚠️ 输出自检清单(强制,每次输出前逐条核对)\r
\r
输出报告前,**必须逐条检查以下项目**,全部通过才可发送:\r
\r
- [ ] **标题链接**:`# [Project Name](GitHub URL)` 格式,可点击跳转\r
- [ ] **标题空行**:每个粗体标题(`**🎯 ...**`)前后各有一个空行\r
- [ ] **Telegram 空行**:每个列表块末尾与下一个标题之间有盲文空格 `⠀` 行(防止 Telegram 吞空行)\r
- [ ] **Issue 链接**:精选 Issue 每条都有完整 `[#号 标题](完整URL)` 格式\r
- [ ] **竞品链接**:每个竞品都附 `[名称](GitHub/官网链接)`\r
- [ ] **社区声量链接**:每条引用都有 `[来源: 标题](URL)` 格式\r
- [ ] **无空泛描述**:社区声量部分没有"评价很高"、"热度很高"等概括性描述\r
- [ ] **信息溯源**:所有外部引用都附原始链接\r
\r
## Dependencies\r
\r
本 Skill 依赖以下 OpenClaw 工具和 Skills:\r
\r
| 依赖 | 类型 | 用途 |\r
|------|------|------|\r
| `web_search` | 内置工具 | Brave Search 检索 |\r
| `web_fetch` | 内置工具 | 网页内容抓取 |\r
| `browser` | 内置工具 | 动态页面渲染(备选) |\r
| `search-layer` | Skill | 多源搜索 + 意图感知评分(Brave + Exa + Tavily + Grok),v2.1 支持 `--intent` / `--queries` / `--freshness` |\r
| `content-extract` | Skill | 高保真内容提取(反爬站点降级方案) |\r
安全使用建议
Before installing or using this skill: 1) Ask the author to update the skill metadata to declare the required credential (e.g., GITHUB_TOKEN or GITHUB_PAT) and to document the exact env var name and required token scope. 2) Do not provide a full-privilege personal token; if you must supply a token, create an ephemeral or least-privilege token limited to read-only access to public repos. 3) Review the external repos (blessonism/openclaw-search-skills, search-layer, content-extract, mineru-extract) before installing — these will run code and may contact third-party services (MinerU). 4) Confirm what TOOLS.md is supposed to contain (SKILL.md references it but none is included). 5) If you cannot review those repos, run the skill in an isolated environment or decline to install. 6) Consider requesting the author to: a) remove hardcoded token examples and b) add explicit declarations of required env vars, third-party endpoints, and data handling (where scraped data and logs are sent). These steps reduce the risk of accidental token exposure or unreviewed code execution.
功能分析
Type: OpenClaw Skill
Name: github-explorer
Version: 1.0.0
The github-explorer skill is a legitimate tool designed to perform deep-dive analysis of GitHub projects by aggregating data from the GitHub API, web searches, and community platforms. It utilizes standard OpenClaw tools and external scripts (search-layer, content-extract) to fetch repository metadata, issues, and social media discussions. While it instructs the agent to use a GitHub Personal Access Token (PAT) via curl and execute local Python scripts, these actions are strictly aligned with its stated purpose and lack any indicators of malicious intent, data exfiltration, or harmful prompt injection.
能力评估
Purpose & Capability
The skill claims to deep-dive GitHub projects, which reasonably requires GitHub API access and multi-source scraping. However, SKILL.md explicitly instructs use of a Personal Access Token (PAT) for GitHub API calls and to run external tools (search-layer, content-extract), yet the skill metadata lists no required environment variables, no primary credential, and no install spec. The missing declaration of the PAT and the loose dependency requirements are disproportionate to the provided registry metadata.
Instruction Scope
Runtime instructions tell the agent to: run curl commands including Authorization: token {PAT}; execute python scripts from search-layer and content-extract; prefer GitHub API over web_fetch; and escalate to MinerU for content extraction. The SKILL.md references a TOOLS.md for the PAT but no such file is included. These instructions therefore require access to secrets (PAT) and call external services/scripts not declared in metadata — the agent would need to read/provide a token and may transmit requests to third-party endpoints.
Install Mechanism
This skill is instruction-only (no install spec) which reduces direct install risk. However README recommends installing or linking third-party repos (blessonism/openclaw-search-skills and its search-layer/content-extract/mineru-extract). Those are external codebases; installing them would execute arbitrary code on the host. The absence of an install spec in the registry means the skill relies on user action to add those dependencies — verify those repos before installing.
Credentials
SKILL.md requires a GitHub PAT for API calls (curl examples use Authorization header), but the registry declares no required env vars or primary credential. content-extract/mineru may also require external API keys (not declared). Asking for a PAT is proportionate to the purpose only if the token requirement and scope (read-only, public repo) are declared; currently they are not, which is a significant omission and a potential exfiltration risk.
Persistence & Privilege
The skill does not request always:true, does not include code files, and does not modify other skills' configs. It appears not to request permanent elevated presence. Autonomous invocation is allowed (platform default) but does not by itself increase concern here.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install github-explorer - 安装完成后,直接呼叫该 Skill 的名称或使用
/github-explorer触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
v1.0.0 - First release: analyze GitHub repositories with deep-dive reports
元数据
常见问题
GitHub项目分析助手 是什么?
Deep-dive analysis of GitHub projects. Use when the user mentions a GitHub repo/project name and wants to understand it — triggered by phrases like "帮我看看这个项目... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 153 次。
如何安装 GitHub项目分析助手?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install github-explorer」即可一键安装,无需额外配置。
GitHub项目分析助手 是免费的吗?
是的,GitHub项目分析助手 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
GitHub项目分析助手 支持哪些平台?
GitHub项目分析助手 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 GitHub项目分析助手?
由 flyingants(@tom859174-sketch)开发并维护,当前版本 v1.0.0。
推荐 Skills