AgentSkillsCN

clean_link_titles

从Markdown文件的链接标题中移除多余信息(如来源、作者等)。

SKILL.md
--- frontmatter
name: clean_link_titles
description: Removes extraneous information (source, author, etc.) from link titles in markdown files.

Clean Link Titles

This skill cleans up the titles of markdown links by removing common suffixes that indicate the source, author, or site name.

Inputs

  • file_path: The absolute path to the markdown file to process.

Supported Patterns (Conservative)

It removes only high-confidence patterns:

Suffixes:

  • | SiteName (e.g., | TechNews)
  • - Source (Only if source contains "News", "新聞", "網", ".com" etc.)
  • _SourceName (e.g., _财经头条)

Prefixes:

  • 【Source】 (e.g., 【TechNews】)
  • [Source] (Length limited to < 15 chars to avoid removing content tags)

Usage

Run the following command in the terminal: python3 .agent/skills/clean_link_titles/clean_titles.py "{file_path}"

Example

Input: - [Title | TechNews 科技新報](...) - [【獨家】Title](...)

Output: - [Title](...) - [Title](...)