Hs Mkt Blog Draft

/blog-draft — Draft a Blog from Captures

Pick up stashed captures and synthesize them into a v1 blog post.

Process

1. Identify source captures. Based on $ARGUMENTS:

   • If specific file paths given: read those captures
   • If a topic/keyword given: search ~/content/captures/ for matching files by tags, title, or content (skip files in captures/done/)
   • If no arguments: list all available captures with their one-line summaries and ask the user to pick

2. Read all source captures thoroughly. Understand:

   • What the insights are
   • How they connect to each other
   • What the narrative arc could be
   • Who the audience is

3. Write a blog draft to ~/content/drafts/<slug>.md with frontmatter:

   ---
   title: "<title>"
   date: YYYY-MM-DD
   status: draft
   source_captures:
     - <path to capture 1>
     - <path to capture 2>
   tags: [<merged from source captures>]
   ---
   

4. Structure the draft. Include:

   • A hook that makes someone want to keep reading
   • Section-by-section content with the actual substance (not placeholders)
   • Code examples where relevant (real, not pseudo)
   • A tone section at the bottom (notes for the human editor on voice, audience, what to emphasize)
   • A sources section referencing the original captures

5. Move source captures from ~/content/captures/ to ~/content/captures/done/. This keeps the active captures directory clean.

6. Report — show the draft path, word count, and the sections written.

Rules

• Write a REAL first draft, not an outline. The user should be able to edit this into a publishable post, not write it from scratch.
• Practitioner voice. Show real commands and code. Developers trust concrete over abstract.
• Don't oversell. "This saved me 20 minutes per session" beats "this 10x'd my productivity."
• The draft should stand alone — a reader shouldn't need to know about the captures to understand the blog.
• Use extended thinking for narrative synthesis.