Brainstorm LinkedIn Posts from Source Material

Generate LinkedIn post ideas based on external context provided by the user.

Process

1. Read ALL published posts (MANDATORY - to avoid topic/angle overlap):

   ./scripts/print-published.sh linkedin-post
   

   This prints all published posts with full content in one call. Note:

   • Core insights already covered
   • Data points already used
   • Angles already explored

   Do not suggest ideas that repeat existing coverage.

2. Read the style guide:

   • guidelines/linkedin.md
   • references/professional-profile.md

3. Process the provided context (user will provide one or more of):

   • Files (PDFs, documents, research papers)
   • URLs (articles, blog posts, announcements)
   • Raw text or ideas

4. Identify 2-4 promising angles by considering:

   • What's the unique insight for a professional audience?
   • How does this connect to the author's expertise?
   • What's the hook that works in the first 210 characters?
   • Is there a personal angle or company connection?

5. Present ideas with for each:

   • Proposed title
   • Core insight (1 sentence)
   • Hook approach (personal anecdote, company experience, surprising outcome, or news)
   • Why it fits the author's voice

6. Ask user to choose:

   • Which idea(s) to develop
   • Whether to create as idea (01-ideas) or draft (02-drafts)

Evaluation Criteria

Strong LinkedIn post ideas have:

• A concrete hook in the first 210 characters
• A clear insight that provides value
• Connection to the author's expertise areas
• Room for a personal or company angle
• Appropriate scope for the target word count

Creating Files

After user selection, get timestamp:

date -u +"%Y%m%d-%H%M%S"  # For slug
date -u +"%Y-%m-%dT%H:%M:%S.000Z"  # For created/lastUpdated

Create file at: content/posts/linkedin-post/{slug}-{slugified-title}.yaml