agentleFS
AI-AnswerAI-Answer/ContentRepurposeSystem/skills/outlines/SKILL.md

outlines

Skill26 starsChanged 16 days agoMITView on GitHub

Generate full YouTube video outlines in your format — hook, structured body with talking points, visual aid callouts, source citations, time budgets, and production notes. Trigger on phrases like 'outline this video', 'write an outline for', 'structure this video', 'create a video outline', 'help me outline', or any request to plan or structure a YouTube video for recording.

6 other projects use a version of this, 2 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create a full video plan with talking points

    line 15
  2. Very specific:

    Always check the vault for related research: search research/ and projects/ for files on the topic.

    line 32
  3. Distinctive:

    Must be fair, give clear recommendation

    line 73
  4. Distinctive:

    Use "head fakes" / misdirection to delay payoff where it serves the story

    line 165
  5. Distinctive:

    If outline runs long, cut the weakest section rather than rushing everything.

    line 183
  6. Distinctive:

    Add 2-3 entries to Production Notes.

    line 193
  7. Distinctive:

    Every section earns its spot: If it doesn't advance understanding or open/close a curiosity loop, cut it.

    line 204

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

YouTube Video Outline Generator

Generate recording-ready video outlines in your exact format. Includes hooks, structured body sections with talking points, visual aid placeholders, source citations, time budgets, competitive differentiation, and production notes.

When This Skill Activates

Activate when the user wants to:

  • Structure a video for recording
  • Turn research or an idea brief into a filmable outline
  • Create a full video plan with talking points
  • Organize raw notes into a video format

Example triggers:

  • "Outline this video about RAG concepts"
  • "Turn this research into a video outline"
  • "Write the outline for tomorrow's video"
  • /outlines or /outlines <topic>

Input

The user provides one or more of:

  • Topic or idea brief — from /ideation or described verbally
  • Research notes — vault files, brain dumps, or pasted research
  • Hook — from /hooks (optional — will generate one if not provided)
  • Target length — how long the video should be (default: 12-18 minutes)

Always check the vault for related research: search research/ and projects/ for files on the topic.

Workflow

Step 1 — Determine Video Format

Classify the video structure based on the content:

Concept/Educational — explaining ideas with escalating depth

  • Structure: Hook → Concept 1 → Concept 2 → ... → Big Picture Closer
  • Example: "7 RAG Concepts Every Claude Code User Should Know"
  • Per concept: core idea (one-liner) → explanation → analogy → source

Level-Based — progressive mastery/identity arc

  • Structure: Hook → Level 1 → Level 2 → ... → Master Level
  • Example: "6 Levels of Claude Code"
  • Per level: what it is → the shift → narrative arc (early/mid/late) → demos → key mantra → transition

Tool/Product Demo — showing how something works

  • Structure: Hook → Problem/Context → Tool Intro → Demo → Honest Limitations → CTA
  • Example: "This Free Tool Replaced My $200/month Software"
  • Heavy on screen recording, light on talking head

Contrarian/Opinion — challenging conventional wisdom

  • Structure: Hook (contrarian claim) → Evidence → Counter-arguments → Resolution → Implications
  • Example: "CLAUDE.md Is Ruining Your Code"
  • Needs strong authority right after the claim

Case Study — documenting a specific result or process

  • Structure: Hook (result reveal) → Context → Process → Results → Lessons → CTA
  • Example: "How I Automated 90% of Social Media with Claude Code"
  • Heavy on demos and proof

Listicle — curated collection

  • Structure: Hook → Item 1 → Item 2 → ... → Best/Most Important Last
  • Example: "5 Open Source AI Projects You Need to Know"
  • Per item: what it is → why it matters → demo moment → who it's for

Head-to-Head — comparing two approaches

  • Structure: Hook → Contender A → Contender B → Direct Comparison → Verdict
  • Example: "RAG vs Agentic Search"
  • Must be fair, give clear recommendation

Step 2 — Build the Outline

# [Video Title] — Video Outline

- **Date:** YYYY-MM-DD
- **Target publish:** YYYY-MM-DD
- **Format:** [Format type]
- **Target length:** [X-Y minutes]
- **Related:** [[wikilink]], [[wikilink]]

---

## Hook (~30-90 seconds)

[If hook provided from /hooks, insert with Three Hook Alignment]

[If no hook, generate one:]
- **Desire mapping:** [Core Desire] → [Proxy Desire]
- **Spoken hook:** "[The actual words]"
- **Visual hook:** [What's on screen]
- **Text overlay:** "[The text]"

---

## [Section 1 Title] (~X minutes)

**The core idea:** [One sentence — what the viewer should understand]

**Talking points:**
- [Specific point with data/example]
- [Specific point with data/example]
- [Transition to next section]

**Visual aid:** → [excalidraw or screen recording description]

**Demo:** [What to show on screen, if applicable]

**Source:** [Citation with URL]

---

## [Section 2 Title] (~X minutes)
[Same structure...]

---

## Outro / CTA (~30-60 seconds)

**Recap:** [One-line per section — rapid fire]

**CTA:** [Subscribe, comment, community link]

---

## Production Notes

### Competitive Landscape
- [Creator] — [Video title] ([X views]) — [How ours differs]
- [Creator] — [Video title] ([X views]) — [How ours differs]

### Title Ideas
1. [Option 1]
2. [Option 2]
3. [Option 3]

### Thumbnail Moments
- [Section where a good thumbnail frame could come from]
- [Key visual that would work as a thumbnail]

### Honest Limitations
- [What the video doesn't cover or where advice breaks down]

### Total Estimated Length: [X minutes]

Step 3 — Apply Narrative Principles

While building, apply the Kallaway dopamine ladder:

  1. Stimulation — the hook (visual stun gun, first 1-2 seconds)
  2. Captivation — plant the core curiosity question early
  3. Anticipation — build toward the key reveal (don't give the best insight first)
  4. Validation — deliver the payoff (must be non-obvious)
  5. Affection — honest limitations, personal experience (likability)
  6. Revelation — big-picture closer that reframes everything

Curiosity loops per section:

  • Each section opens a mini-loop (question/tension) and closes it
  • Best sections close one loop while opening the next
  • Use "head fakes" / misdirection to delay payoff where it serves the story

Law of Interestingness:

  • Front-load the most surprising facts (shock score 70+)
  • Broadest context first (the "bedrock")
  • Most perspective-shifting insight saved for the ending

Fifth-grade reading level for spoken content. Short sentences. Common words. Define technical terms immediately.

Step 4 — Time Budget

Assign time estimates per section. Total must match target length.

  • Hook: 30-90 seconds (never longer)
  • Individual section: 2-4 minutes each
  • Demo segments: 3-7 minutes
  • Outro: 30-60 seconds

If outline runs long, cut the weakest section rather than rushing everything.

Step 5 — Competitive Check

Optional step. This uses a separate yt-search helper script that does not ship with this plugin. If it isn't installed, fill in the competitive landscape from what you already know about the space (or skip it) — the outline is complete without it.

python "${CLAUDE_PLUGIN_ROOT}/skills/yt-search/scripts/search.py" <topic> --count 10 --months 3

Add 2-3 entries to Production Notes. For each competitor: their angle, performance, how yours differs.

Step 6 — Save to Vault

Save to YOUR_VAULT_PATH/projects/YYYY-MM-DD-<slug>-video.md.

Link to research files and idea briefs with wiki links.

Key Principles

  • Outlines are for speaking, not reading. Talking points are things you would say out loud. No academic language.
  • Every section earns its spot. If it doesn't advance understanding or open/close a curiosity loop, cut it.
  • Visual aids are not optional. Flag every section needing a diagram, screen recording, or visual. No talking head longer than 60 seconds without a visual change.
  • Sources build credibility. Specific numbers, papers, people, tools. "Studies show..." is worse than no citation.
  • Honest limitations build trust. At least one "here's where this breaks down" moment per video.
  • Production notes are for you. Title ideas, thumbnail moments, competitive landscape — filming and publishing decisions.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.