Create Blog Series

Do NOT attempt to read file contents — the blog_writer subagent handles all parsing.

Phase 2: Generate

Spawn the blog_writer subagent using the Task tool with only:

• Folder path: Full path to the research output folder (e.g., research_output/framework-claude-agent-sdk/)
• Today's date: For frontmatter date fields (today is {{date}})

The blog_writer will independently:

1. Read and parse all source files (README.md, learning-path.md, etc.)
2. Discover optional content (practical-takeaways.md, code-examples/)
3. Create the directory structure under blog/{series-slug}/
4. Generate all blog posts with proper frontmatter
5. Wire prev/next navigation links across subfolders

Phase 3: Verify

After the blog_writer completes, use Bash to verify the output:

1. Run ls -R research_output/{folder}/blog/ to confirm the file structure
2. Report to the user showing:
   • Total parts generated
   • File paths for each part
   • Which optional parts were included/skipped

Output Structure

research_output/{topic}/blog/{series-slug}/
├── 00-introduction.md
├── chapters/
│   ├── 01-{level-1-slug}.md
│   ├── 02-{level-2-slug}.md
│   ├── 03-{level-3-slug}.md
│   ├── 04-{level-4-slug}.md
│   ├── 05-{level-5-slug}.md
│   └── 06-practical-takeaways.md         ← if practical-takeaways.md exists
└── code-examples/
    └── 07-hands-on-code.md               ← if code-examples/ exists