pull

Get Repository Info

Fetch all public, non-fork (source) repositories from a GitHub account and create update.md in the repository root with categorized entries when new repos exist.

When to Use

• Cataloging new repos that need to be added to this repository
• Regenerating the update.md file with fresh data from GitHub

Script Reference

See scripts/README.md for script details and outputs.

Procedure

1. Run the fetch script:

• Run from the workspace root:

  python .github/skills/pull/scripts/fetch_repos.py
  

• Use --dry-run to list new repos without fetching or writing files.
• Review the script output for the summary.
• If the script fails, reports 0 new repos, or every repo has ERROR: markers, stop and ask before continuing.

2. Classify and categorize each repo in update.md:

• For each repo, inspect the README: field first.
• If README: points to a file path, read that README and check the metadata (Language, Topics, Homepage).
• If README: is ERROR:, classify using metadata only. If the category is unclear, stop and ask before continuing.
• If it is NOT a code repo (e.g., website/blog, GitHub Pages, Hugo content-only site, profile README, slides without code):
  • Add the repo name to exclude-repos.md on its own line.
  • Remove it from update.md.
• If it IS a code repo, assign a category:
  • Check existing categories in README.md (## section headings).
  • Use an existing category if it fits, otherwise create a new one.
• Pay attention to repos with ERROR: markers — they may need manual review.

3. Rewrite update.md with categorized entries:

   # Category Name
   
   ## repo-name
   https://github.com/parsiya/repo-name
   
   Description from GitHub.
   
   blog link: URL (if found in repo README or Homepage field)
   blog title: Title (if found in repo README)
   

• Group repos by category (# heading) with each repo as ##.
• If you are creating a new category, clearly mention it.
• Include the GitHub URL, description, and any blog links found in the repo's README or Homepage field.
• Remove the metadata fields (Language, Topics, README path) — those were just for classification.

4. Report summary: Tell the user how many repos were found, how many were filtered out, and how many are in update.md.