Introduction

To quickly find out which GitHub projects are gaining stars today, this week, or this month, common practices include opening github.com/trending and browsing the list manually, or writing a web crawler to parse the HTML into a list. The former is time-consuming, while the latter requires maintaining parsing logic and an additional round of effort to organize the “rankings” and explain “why they are worth watching.”

Here we introduce github-trending on SkillHub (coordinates @user_cea1fb81/github-trending-pro). It integrates fetching the Trending page, structured output, and report-style analysis into the Agent’s SKILL workflow: when a user requests a hot list or trend analysis in natural language, the Agent calls the script according to the skill instructions to fetch data, writes a readable report based on a fixed format, and is constrained not to fabricate fields such as stars or descriptions.

What is This

github-trending is a community Skill for coding Agents (Source: community, Category: data-analysis, Version: 1.0.0). It is published by SkillHub user user_cea1fb81 (displayed as “Running Crispy Sausage”).

The skill’s positioning is described directly in the official introduction: it is used when users ask about GitHub Trending, popular repositories, daily/weekly/monthly rankings, filtering by programming language, filtering by spoken language, or when performing trend analysis on popular projects. SkillHub tags indicate requires_api_key=false, meaning it does not rely on third-party API Keys; data is obtained via HTTP scraping and HTML parsing of the public GitHub Trending page.

Core Functions

Upon verification of SKILL.md and scripts/fetch_trending.py within the skill package, the confirmed capabilities are as follows.

List Scraping and Filtering

  1. Time Range: daily, weekly, monthly; default is daily.
  2. Programming Language: Corresponds to the GitHub Trending language path, e.g., python, rust, javascript; if not specified, the all-languages page is used.
  3. Spoken/Documentation Language: Passed via --spoken-language using a language code (e.g., zh, en), which maps to the spoken_language_code parameter in the Trending URL.
  4. Quantity: --limit controls the number of returned repositories; the skill instructions default to Top 10.

The target URL looks like https://github.com/trending or https://github.com/trending/{language}?since=..., constructed by the script’s build_url function.

Structured Output

The script outputs JSON by default, containing report metadata (date, since, language, spoken_language, url, updated_at) and a repository list. Each repository includes both compatibility fields (rank, repo, url, description, language, total_stars, forks, current_period_stars) and extended fields for easier analysis (full_name, author, summary, primary_lang, etc.).

When a user explicitly requests a table or export, --format markdown can be used to generate Markdown with an Overview table.

Analysis Report (Agent Layer)

The skill requires the Agent to do more than just copy the list: it must write Feature / Functionality / Implementation / Why Trending based on the JSON facts, and perform thematic summarization (e.g., AI agent, developer tools, infrastructure). The analysis must be based on script output and public descriptions; if information is missing, state “not specified” or omit it, and do not fabricate stars, URLs, or author information.

Installation and Activation

Install SkillHub CLI

If skillhub is not installed on your machine, you can check or install it according to the official installation documentation:

command -v skillhub && skillhub --version

If not installed, a full installation (CLI + default Skills) example:

curl -fsSL https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/install/install.sh | bash

See SkillHub Installation and Priority Source Configuration for instructions.

During installation, you must use --dir to point to the current Agent’s skills directory; otherwise, the skill may land in ./skills/ in the current directory and fail to be loaded by the Agent. The installation command provided by official search is:

skillhub install github-trending-pro --namespace user_cea1fb81 --dir <skills directory>

Common paths (from SkillHub installation documentation):

  • Cursor: ~/.cursor/skills/
  • Claude Code: ~/.claude/skills/
  • Codex: ~/.codex/skills/ or .agents/skills/ under the project

Alternatively, you can ask the AI in the conversation to install @user_cea1fb81/github-trending-pro following https://skillhub.cn/install/skillhub.md.

After installation, restart or refresh the Agent as required by your client so the skill is discovered.

Runtime Environment

The scraping script is Python 3 and uses only the standard library (urllib, html.parser, etc.), requiring no additional pip dependencies. It needs access to github.com during execution (the script’s request headers include User-Agent: codex-github-trending-skill/1.0).

Typical Usage

Trigger in Conversation

Simply describe your needs in natural language, aligning with the skill’s description, for example:

  • “List the top 10 GitHub Trending today and analyze AI-related themes.”
  • “What are the trending Rust projects this week?”
  • “Show popular JavaScript repositories from the monthly list, with a preference for Chinese documentation language.”

The Agent should infer since, language, and spoken-language, execute scripts/fetch_trending.py in the skill directory, and use JSON for analysis by default; Markdown is only selected when the user requests a raw table.

Run Script Directly (Reproducible)

Execute within the installed skill directory (path depends on your local installation):

python3 scripts/fetch_trending.py --since daily --limit 10

Filter by language and time period, outputting JSON:

python3 scripts/fetch_trending.py --language python --since weekly --limit 10 --format json

Example with spoken language filter (command from skill documentation):

python3 scripts/fetch_trending.py --language rust --since monthly --spoken-language zh --limit 10

For a Markdown overview table:

python3 scripts/fetch_trending.py --language javascript --since daily --format markdown

Report Structure

For human-readable reports, the skill requires a title, date, range, language, and spoken language filter first, followed by an Overview with fixed headers, for example:

| # | Repo | Total Stars | Today | Lang | Description |
|---|------|-------------|-------|------|-------------|

For weekly or monthly lists, the “Today” column name can be changed to This Week or This Month (unless the user explicitly requests keeping “Today”). Details for each repository are written according to the skill template covering URL, stars, growth, language, and segmented analysis, ending with a brief trend summary.

Use Cases and Notes

Target Audience: Developers or technical writers who need to regularly scan GitHub hot lists, narrow down scope by language or documentation language, and want the Agent to provide a layer of summarization based on factual data; categorized as data-analysis, focusing on “data retrieval + interpretation” rather than simple searching.

Limitations and Notes:

  1. Page Structure Dependency: The parsing logic targets GitHub Trending HTML. If GitHub changes its layout, the list may be empty or fields may be missing; the script outputs JSON-formatted error to stderr upon failure.
  2. Network and Compliance: Scraping is affected by local network and GitHub accessibility; please comply with GitHub’s Terms of Service and use reasonable access frequencies.
  3. Analysis Boundary: Sections like Implementation and Why Trending are inferences by the Agent under constraints and should be distinguished from raw descriptions in the script; do not mix prose when the user requests pure JSON.
  4. Store Data: As of the verification of this material, SkillHub shows a rating of 100000, downloads of 974, and installs of 1; refer to the Skill Page for real-time data.

Conclusion

github-trending connects the filtering dimensions of GitHub Trending (daily/weekly/monthly, language, spoken language) and a verifiable JSON list to the Agent’s reporting process, reducing the time spent manually opening web pages and copying tables. If you want to use it regularly in environments like Cursor, simply install it to the corresponding skills directory using skillhub install as described above.

Official Entry Points:

  • SkillHub Skill Page: https://www.skillhub.cn/skills/user_cea1fb81/github-trending-pro
  • API Coordinate Page: https://api.skillhub.cn/user_cea1fb81/github-trending-pro