Unlike generic prompts, NovaKit skills check what is working in your category before generating. Here is what that produces across different use cases.
💼
LinkedIn Post Engine
Posts calibrated to this week
Checks what hook formats, post lengths, and topic angles are earning reach right now — not what worked six months ago. Run the same skill twice on the same topic and get different posts, because the context will have shifted.
→ Consistent reach without guessing what the algorithm wants
📊
Pitch Deck Narrative
Investor narrative, not slide bullets
Asks 3 questions about your business, then structures a narrative around the objections your specific investor type actually raises — not a generic problem/solution/traction template every founder uses.
→ Pitch structure that handles the questions before they're asked
📬
Cold Outreach Engine
Messages that get replies
Researches your target's recent activity and what's getting responses in their industry before writing the message. The result is specific enough that recipients know it wasn't mass-sent — without you doing the research manually.
→ Reply rates that hold across multiple send batches
📝
PRD Writer
Requirements that shape what gets built
Asks about the decision that kicked off the PRD, the constraints the team is working within, and who has approval authority — then writes requirements that reflect those realities, not a generic feature spec.
→ PRDs engineers actually use to make tradeoff decisions