Podcast Episode Planner Skill

Purpose: Transform research materials into a structured episode plan that guides NotebookLM's two-host audio generation.

When to Use: After research is complete (Phase 7 report.md exists) and before audio generation. This skill creates content_plan.md - the structural blueprint that NotebookLM uses to create coherent, well-organized podcast audio.

Skill Overview

This skill produces a structured episode plan that:

1. •Classifies the episode type and selects appropriate toolkit elements
2. •Structures content across three sections (Foundation, Evidence, Application)
3. •Identifies key terms, studies, and narrative elements to emphasize
4. •Provides guidance for NotebookLM's two-host conversation format

Key Principles:

• •Plans provide structural guidance for NotebookLM, not verbatim scripts
• •Section 1 introduces concepts and methodology — conclusions come later
• •Acronyms must be spelled out on first use ("polyunsaturated fatty acids, or PUFA")
• •Use metric units or intuitive measures ("handful", "palm-sized") — not imperial

Inputs Required

Before invoking this skill, gather:

1. Episode Metadata

topic: "[Specific topic/angle]"
core_question: "[The single question this episode answers]"
title: "[Episode title]"

2. Series Context (if applicable)

series_name: "[Series name]"
series_question: "[Core question the series answers]"
position: "[opener | middle | closer | standalone]"
episode_number: [N]

3. Source Material (from report.md and p3-briefing.md)

key_studies:
  - name: "[Study name/author/year]"
    finding: "[Key finding]"
    strength: "[Meta-analysis | RCT | Observational | etc]"

surprising_findings:
  - "[Finding that challenges assumptions]"

contradictions:
  - topic: "[Where sources disagree]"
    resolution: "[How to present this]"

practical_protocols:
  - "[Actionable insight with specific parameters]"

Planning Process

Step 1: Episode Classification

## Episode Classification

- **Series Position:** [opener / middle / closer / standalone]
- **Evidence Status:** [consensus / minor conflict / major conflict]
- **Content Density:** [concept-heavy / protocol-heavy / balanced]

Step 2: Toolkit Selection

Opening Hook (choose ONE):

Takeaway Structure (choose ONE):

Contradiction Handling (if evidence contested):

Step 3: Section Planning

Section 1: Foundation

Focus: WHY - Introduce concepts and research methodology. Save conclusions for later.

Plan:

• •Opening hook content
• •Key concepts to introduce (2-3 max)
• •Terms that MUST be defined (with pronunciation if unusual)
• •Analogies to anchor abstract concepts
• •Transition to Section 2

Section 2: Evidence

Focus: WHAT - Present the research findings.

Plan:

• •Key studies to highlight (with institution, year, sample size)
• •Evidence clusters (group related findings)
• •Where evidence agrees vs. conflicts
• •Callbacks to Section 1 concepts
• •Transition to Section 3

Section 3: Application

Focus: HOW - Translate findings to action.

Plan:

• •Specific protocols with parameters (timing, frequency, dosage)
• •Who this applies to / caveats
• •Callback to opening hook (complete the arc)
• •Final synthesis

Step 4: State Tracking

Track what's been established to enable callbacks:

• •Terms defined: [list - can use freely after definition]
• •Concepts established: [list - can callback without re-explaining]
• •Open loops: [questions raised that must be answered by end]

Step 5: Quality Check

Before finalizing, verify:

• • Three sections with clear focus (Foundation/Evidence/Application)
• • Maximum 3-4 major concepts per section
• • All key terms listed with definitions
• • Protocols include specific parameters
• • Opening hook connects to closing callback
• • Episode answers its stated core question

Output Format

Produce content_plan.md in this structure:

# Episode Plan: [Episode Title]

## Episode Metadata
- **Series:** [Series name or "Standalone"]
- **Position:** [Opener / Middle / Closer / Standalone]
- **Core Question:** [The question this episode answers]
- **Episode Type:** [Evidence status] + [Content density]

## Toolkit Selections
- **Hook Type:** [Selected hook]
- **Takeaway Structure:** [Selected structure]
- **Contradiction Handling:** [Approach if applicable]

---

## NotebookLM Guidance

### Opening Instructions
[Specific guidance for how the hosts should open - the hook to use, tone to set]

### Key Terms to Define
| Term | Definition | Pronunciation (if needed) |
|------|------------|---------------------------|
| [Term 1] | [Clear definition] | [e.g., "poo-fah" for PUFA] |
| [Term 2] | [Clear definition] | |

### Studies to Emphasize
1. **[Study name, Institution, Year]** - [Key finding to highlight]
   - Sample size: [N]
   - Why it matters: [Context]

2. **[Study name]** - [Key finding]

### Narrative Arc

**Section 1: Foundation**
- Primary focus: [What concept/mechanism to establish]
- Key analogy: "[Everyday comparison to anchor the concept]"
- Transition hook: [How to move to evidence]

**Section 2: Evidence**
- Evidence cluster A: [Studies supporting point 1]
- Evidence cluster B: [Studies supporting point 2]
- Conflict to address: [If any, how to present it]
- Callback opportunity: "[Reference to Section 1 concept]"

**Section 3: Application**
- Protocol 1: [Specific action with parameters]
  - Timing: [Specific]
  - Frequency: [Specific]
  - Who: [Target population]
- Protocol 2: [If applicable]
- Caveats: [Important limitations]

### Closing Instructions
- Callback to opening: [How to reference the opening hook]
- Key takeaway: [Single most important point]
- Sign-off: "Find the full research and sources at research dot yuda dot me—that's Y-U-D-A dot M-E."

---

## Specificity Standards

The hosts should use specific parameters throughout:

| Category | Vague (Avoid) | Specific (Use) |
|----------|---------------|----------------|
| Timing | "in the morning" | "90-120 minutes after waking" |
| Frequency | "regularly" | "3 times per week" |
| Citations | "some studies show" | "A 2023 meta-analysis of 47 trials found" |
| Effects | "significant improvement" | "17% reduction in all-cause mortality" |
| Dosage | "take some magnesium" | "300-400mg magnesium glycinate" |

---

## Attention Maintenance Notes

Remind hosts to:
- Rotate content types every 5-7 minutes (explanation → example → insight)
- Use pattern interrupts every 7-10 minutes
- Signpost major transitions ("Key point here...", "This brings us to...")
- Close any open loops before episode end

Templates

Opening Hook Templates

Provocative Question:

"What if everything you believe about [topic] is fundamentally wrong?"

Surprising Statistic:

"[Specific number] [unexpected comparison]. That's [X times more/less] than [common assumption]."

Bold Claim:

"By the end of this episode, you'll understand exactly how to [specific outcome]."

Counterintuitive Claim:

"The experts have been wrong about this for decades. And the data finally shows us why."

Stakes Establishment:

"This single [factor] predicts [outcome] better than any other—and most people are getting it completely wrong."

Series Modifier Templates

Series Frame Opening:

"This is the [ordinal] episode in our series on [topic]. Today, we're looking at it through the lens of [perspective]."

Series Wrap (closer):

"This concludes our series on [topic]. Together, these perspectives give you a complete framework for [core question]."

Callback Templates

• •"As we covered earlier, [concept]—this is exactly why [new point]."
• •"Remember the mechanism we discussed? This study shows it in action."
• •"This brings us back to [opening hook reference]. Now you understand why."

Integration with NotebookLM

The content_plan.md file is uploaded to NotebookLM along with:

• •research/p1-brief.md - Research brief
• •report.md - Narrative synthesis
• •research/p3-briefing.md - Master briefing
• •sources.md - Validated citations

NotebookLM uses content_plan.md to:

1. •Structure the conversation flow
2. •Know which terms to define and when
3. •Emphasize the right studies
4. •Create coherent narrative arc with callbacks
5. •Deliver specific, actionable protocols

The episodeFocus prompt in the NotebookLM API provides additional guidance on tone and style.

Skill Version: 3.0 Output: content_plan.md only (NotebookLM guidance)