develop

/develop — Develop Content Concept

Take a content idea stub and develop it into a fully strategized, packaged concept.

Variables

$ARGUMENTS (the stub ID like "#5", or a raw idea to capture and develop in one go)

Instructions

You are running the Develop step of the Content Pipeline. Take a captured idea and turn it into a fully developed content concept with strategic positioning, audience alignment, packaging, and offer mapping.

Setup — Load Context

1. Read the idea:

   • If given an ID (#N): query SELECT * FROM content_ideas WHERE id = N via the content database
   • If given raw text: capture it as a stub first (classify channel + format), then develop

2. Read content strategy docs (ALWAYS read all of these):

   • content/strategy.md — platform, cadence, pillars, audience split, competitive positioning
   • content/brand-and-audience.md — brand positioning, target audience segments, proof points
   • content/offers-and-funnels.md — offers, funnels, audience → offer alignment, CTAs
   • content/packaging-strategy.md — title/thumbnail/hook frameworks (used for both YouTube AND LinkedIn)

3. Scan inspiration library (check for relevant examples):

   • content/inspiration/linkedin/ — LinkedIn post examples and hooks worth referencing
   • content/inspiration/youtube/ — YouTube title/thumbnail examples
   • content/inspiration/frameworks/ — posting frameworks and content systems

   If any examples match the concept's channel or topic, reference them in packaging ("This hook pattern is inspired by [source]"). Don't copy — adapt and make it yours.

4. Build the 7-day context window:

   .venv/bin/python3 -c "
   import sys; sys.path.insert(0, '.')
   from scripts.content.context_aggregator import build_context_window, format_context_for_prompt
   context = build_context_window(days=7)
   print(format_context_for_prompt(context))
   "
   

   Read the output — this tells you:

   • Recent published content (what's already been covered so you don't repeat it)
   • Recent meetings (themes, client pain points, insights worth referencing)
   • Current pipeline state (what's already queued)