github-auto-star

Auto Star GitHub Projects

Overview

This skill can automatically extract GitHub repository references from various content sources and star them on your GitHub account.

Supported trigger methods:

"Analyze this article and star the projects in it"
"Find GitHub projects in this screenshot and star them"
"Star all GitHub projects mentioned in [content]"
Set up automated scheduled tasks

Dependencies

System dependencies

Dependencies	Installation method
GitHub CLI (gh)	macOS: `brew install gh`Linux: `sudo apt install gh`

Environment variables

Variable	Description	Configuration method
`GITHUB_TOKEN`	GitHub Personal Access Token (optional, logging in with gh CLI is recommended)	Copy `config/.env.example` to `config/.env` and fill in the actual values

Prerequisites

GitHub authentication must be configured before using this skill. For configuration steps, see github-config.md。

Workflow

Step 1: Content extraction and context analysis

Extract content and context information from the provided source:

Text/URL:

Use WebFetch to retrieve the article content
Parse GitHub URLs (github.com/owner/repo)
Identify project names, library names, and tool names
Extract context: article topic, technical domain, feature description

Screenshots/Images:

Use the image analysis MCP (provided by the user)
Extract text and identify project references
Analyze the image context：
- Identify the technical domain (e.g., databases, AI Agents, frontend frameworks, DevOps, etc.)
- Extract feature descriptions and characteristic keywords
- Understand the project's use case scenarios
- Identify the relevant tech stack

Step 2: Repository discovery and intelligent matching

For each identified project, perform intelligent matching based on context:

Direct match: a complete GitHub URL found in the content, use it directly
Search by name(when only the project name is available):
```
gh search repos "项目名" --limit 10
```
Context relevance verification(key step):

When multiple projects with the same or similar names exist, matching must be based on context:
- Domain match: check whether the repository's topics and description are consistent with the technical domain in the content
  - Example: if the content mentions "AI Agent", prioritize repositories whose topics include ai、agent、llm Example: if the content mentions "AI Agent", prioritize repositories whose topics include ai, agent, llm
  - Example: if the content mentions "database tools", prioritize repositories whose topics include database、sql Example: if the content mentions "database tools", prioritize repositories whose topics include database, sql
- Feature keyword match: compare the repository description with the feature description in the content
  - Extract feature keywords from the content
  - Match them against the repository's description and README
- Tech stack match: Check the repository's primary language and tech stack
  - Language information:gh repo view owner/repo --json primaryLanguage
  - Dependency analysis: check package.json, requirements.txt, etc.
- Popularity and activity reference(as auxiliary):
  - Star count, last update time, Fork count
Matching decision：
- High-confidence match: star directly
- Medium confidence: show the candidate list to the user and request confirmation
- Low confidence: skip and record, mark "needs manual confirmation" in the report

Step 3: Check whether already starred

Before starring, verify whether the repository has already been starred:

gh api user/starred/owner/repo 2>/dev/null && echo "已 star" || echo "未 star"

Step 4: Star the repository

Perform the star operation:

gh repo star owner/repo

Step 5: Generate report

Create a star report, including:

List of newly starred repositories
Already starred repositories (skipped)
Repositories needing manual confirmation (multiple candidates or insufficient confidence)
Repositories not found/failed

Output format

Provide a summary report:

## Star 报告

### 新 Star (3)
- [owner/repo](url) - 描述（匹配依据：领域/功能/技术栈）

### 已 Star (2)
- owner/repo

### 需人工确认 (1)
- 项目名
  - 候选 1: [owner1/repo](url) - 描述 ⭐ 推荐（匹配度：高）
  - 候选 2: [owner2/repo](url) - 描述
  - 上下文提示：内容提到 AI Agent 开发，推荐选择候选 1

### 未找到 (1)
- 项目名（在 GitHub 上找不到匹配项）

Scheduled task configuration

Implement automated execution via hooks, configured in settings.json:

{
  "hooks": {
    "Prompt": [
      {
        "matcher": "star",
        "hooks": ["auto-star"]
      }
    ]
  }
}

Users can configure custom scheduled task sources:

RSS feed URL
Blog URLs to monitor
Execution frequency (e.g., every 2 hours)

Reference materials

GitHub configuration guide - Token setup instructions

Notes

GitHub API limits: about 60 requests/hour unauthenticated, 5000 requests/hour after authenticating via gh
Consider adding delays between star operations
Always verify repository relevance before starring