The Chat-to-Script Extractor

Extract a shareable narrative script and key decisions from a long AI work session — ready for NotebookLM, a slide deck, or a team briefing.


I’m going to give you a long conversation I had while building $ARGUMENTS

If no topic was provided above, ask me what the conversation was about before proceeding.

Read through it completely before responding. Then produce the following:

  1. Core Problem Statement (2-3 sentences): What was I trying to solve, and why did it matter?

  2. Key Decisions (3-5 bullets): The actual choices made — architectural, methodological, strategic. Be specific. Not “we discussed options” but “we chose X over Y because Z.”

  3. The Architecture or Process (1-2 paragraphs): How the solution actually works. Enough detail that someone could implement a version of it.

  4. The Insight Worth Sharing (1 paragraph): The non-obvious discovery or reframe that came out of this work — the thing that would make a colleague say “I hadn’t thought about it that way.”

  5. Narrated Script (400-600 words): Write this as a first-person narration, as if I’m walking someone through a short video presentation of this work. Present tense. Concrete specifics. No filler.

Here is the conversation: [PASTE CHAT HERE]


Source

From 2026-04-23_Mastermind: Lou’s cognitive twin pipeline — the cognitive mirror skill scans a session, extracts decisions and architecture, and produces a narrative script. That script feeds NotebookLM (which generates podcast-style audio) or goes directly into slides. A 6-to-8-hour development session becomes a 6-minute narrated video in roughly 30 minutes. “I now have a process I can use every single time.” The command extracts the pipeline step that turns raw chat into structured, shareable narrative.