Article Processing

2) Duplicate Check

• Search existing notes under ./Articles/**/*.md:
  • rg 'uuid:\s*"?{generated_uuid}"?' ./Articles -g '*.md'
• Match both quoted and unquoted UUID values.
• If a match exists, stop and report the existing note path.

3) Content Extraction

Primary method:

• Use agent-browser skill to open and extract article content.
• Detect Medium paywall indicators such as:
  • Create an account to read
  • author made this story available to
• If paywall is detected, use Freedium mirror:
  • https://freedium-mirror.cfd/<original_url>

Fallback:

• If browser extraction fails, ask user for pasted article content.

4) Content Analysis

Detect source type and focus only on core article content:

• Medium via Freedium:
  • Content: document.querySelector("body > div.container > div")
  • Title: document.querySelector("h1")
  • Author/date block: document.querySelector("body > div.container > div > div:nth-child(2)")
• Medium direct:
  • Content: first <article> element
• Dev.to:
  • Detect by dev.to domain or meta[name="application-name"] == "dev.to"
  • Content: document.querySelector("#article-body")
  • Title: meta[property="og:title"]
  • Author fields from #article-show-container data attributes
  • Date from #main-title time
• Other sites:
  • Use extracted main article body.

Image analysis scope:

• Include only images that belong to article content.
• Ignore nav, ads, social widgets, and unrelated media.
• Handle <img>, <picture> (choose highest-quality source), CSS backgrounds, SVG references.
• Build a complete list of article images before downloading.

5) Image Download

Download order:

• Try aria2c, then wget, then curl.
• Download to temporary tmp_images/ first.
• Retry failed images with the next tool.
• Move successful files to:
  • ./Articles/{main_topic}/{english_article_title}/attachments/

Renaming rules:

• Use short contextual English names.
• Force .jpg extension for JPEG images (never .jpeg).
• Keep other extensions when appropriate (.png, .webp, .svg).

Reference rule:

• Use local links only:
  • ![Alt text](attachments/{image_name}.jpg)

6) Obsidian Content Creation

Main Topic Rules

• Set main_topic as the first non-system tag.
• Exclude article, en, es.
• Normalize to lowercase slug.
• Replace / with - (example: git/worktree -> git-worktree).

Folder and File Structure

• Article folder:
  • ./Articles/{main_topic}/{english_article_title}/
• English note:
  • ./Articles/{main_topic}/{english_article_title}/{english_article_title}.md
• Spanish note:
  • ./Articles/{main_topic}/{english_article_title}/{spanish_article_title}.md
• Attachments:
  • ./Articles/{main_topic}/{english_article_title}/attachments/

Language Rules

• If source is English: create English + Spanish notes.
• If source is Spanish: create Spanish + English notes.
• If source is another language: create only English + Spanish notes (do not keep original-language note).

Filename Rules

• Preserve title text as much as possible.
• Use natural-language filenames (with spaces), not dash-separated slugs.
  • Example: Como uso Claude Code 50 consejos practicos.md (correct)
  • Not: como-uso-claude-code-50-consejos-practicos.md
• Replace invalid macOS filename chars:
  • : -> -
  • / and \\ -> -
  • * -> -
  • ? -> ``
  • " -> '
  • < and > -> ()
  • | -> -

7) Base Compatibility Update

Articles/Articles.base depends on this:

• English note must include:
  • spanish_file: "{spanish_filename}.md"
• spanish_file must exactly match the Spanish filename.
• Keep language: "en" / language: "es" in frontmatter.
• Do not include en or es in tags.

8) Cleanup

• Remove temporary tmp_images/ after files are moved.
• Verify files, links, frontmatter, and attachments are complete.

Frontmatter Template

---