slidev-deck

by Peleke

Install this Skill

curl -fsSL -o slidev-deck.zip https://skill.seembox.com/api/v1/skills/slidev-deck/download && unzip slidev-deck.zip

SKILL.md About

---
name: slidev-deck
description: Generate product-oriented slide decks using sli.dev with embedded hand-drawn SVGs. Use when users ask for pitch decks, strategy presentations, validation summaries, or any visual deck that needs to tell a product/engineering story. Integrates with the hunter pipeline outputs and uses the sketches skill for SVG generation.
metadata:
  openclaw:
    emoji: "\U0001F3AC"
---

# Slidev Deck

Generate complete, ready-to-present slide decks using sli.dev (Markdown-based presentations) with inline hand-drawn SVGs in the portfolio's sketchy style. Each deck tells a visual story -- SVGs carry the narrative, text stays minimal. Decks pull structured data from the hunter product-discovery pipeline and produce a single `slides.md` file runnable with `npx slidev`.

## Step 0: Load Conventions

Before doing ANYTHING, read the shared conventions file:

```
Read /Users/peleke/Documents/Projects/skills/skills/custom/_conventions.md
```

This file defines: canonical vault path, folder-to-type mapping, frontmatter contract, valid statuses, tag hierarchy, cross-reference syntax, and the PipelineEnvelope schema. All output from this skill MUST conform to those conventions. If there is a conflict between this SKILL.md and `_conventions.md`, the conventions file wins.

## When to Use

- Creating a pitch deck for external audiences (warm leads, partners, potential customers)
- Summarizing hunter pipeline validation results for internal review
- Presenting technical architecture to engineering audiences
- Building strategy decks for business positioning and revenue planning
- Any situation where a visual presentation needs to tell a product or engineering story

## Trigger Phrases

- "Build a pitch deck for [product]"
- "Create a validation summary deck from [pipeline outputs]"
- "Make a strategy presentation for [product/domain]"
- "Generate a technical architecture deck for [system]"
- "Slide deck for [topic]"
- "/slidev-deck [type] [product]"

---

## Prerequisites

Before starting, establish:

1. **Deck type** -- One of: `product-pitch`, `validation-summary`, `technical-architecture`, `strategy` (see Deck Types below)
2. **Product/domain** -- What product, system, or domain the deck is about
3. **Audience** -- Who will see this (warm leads, investors, internal team, engineers, etc.)
4. **Pipeline data** -- Which pipeline outputs are available (signal scans, decisions, personas, offers, pitches). Glob the vault to find them:
   ```
   Glob: Admin/Product-Discovery/**/*.md
   ```
5. **Key message** -- The single sentence the audience should remember after the deck

If the user provides only a product name, ask about deck type, audience, and key message before proceeding. If pipeline data exists in the vault, load it automatically.

---

## SVG Style Contract

All SVGs in decks MUST follow the portfolio's dark-theme sketchy style. This is non-negotiable.

### Color Palette (Dark Slide Background)

| Role | Color | Hex |
|------|-------|-----|
| Slide background | GitHub dark | `#0d1117` |
| Primary text | Cream | `#eee0cc` |
| Accent (coral) | Warm coral | `#e88072` |
| Success (green) | Bright green | `#4ade80` |
| Danger (red) | Soft red | `#f87171` |
| Muted | Gray | `#888888` |
| Box fill | Translucent coral | `rgba(232,128,114,0.06)` |
| Highlighted box fill | Translucent coral (stronger) | `rgba(232,128,114,0.12)` |

### Font

```
font-family="'Caveat', 'Comic Sans MS', cursive"
```

- Titles: 20-24px, bold
- Labels: 14-16px
- Annotations: 11-13px
- Captions: 10-12px, muted (`#888888`), reduced opacity

### Sketchy Primitives

**Lines**: Never use `<line>`. Use `<path>` with cubic beziers and 1-3px jitter on control points.

```xml
<path d="M 50,51 C 90,48 130,53 170,49 C 210,52 240,48 250,51"
      stroke="#eee0cc" stroke-width="2" fill="none" stroke-linecap="round"/>
```

**Rectangles**: Never use `<rect>` for content boxes. Build from 4-segment `<path>` with jittered corners.

```xml
<path d="M 132,68 C 230,65 430,71 562,67
         C 565,85 563,105 565,128
         C 430,131 230,127 132,131
         C 129,113 131,87 129,68"
      stroke="#e88072" stroke-width="2" fill="rgba(232,128,114,0.06)"
      stroke-linecap="round" stroke-linejoin="round"/>
```

**Circles**: Use bezier-approximated circles, not `<circle>`, except for numbered callouts.

**Numbered callouts**: Coral fill circle with white bold number text.

```xml
<circle cx="108" cy="82" r="14" fill="#e88072" stroke="none" opacity="0.9"/>
<text x="108" y="87" text-anchor="middle"
      font-family="'Caveat', 'Comic Sans MS', cursive"
      font-size="14" fill="white" font-weight="bold">1</text>
```

**Arrows**: Bezier path body + marker-end arrowhead (or inline chevron path).

```xml
<defs>
  <marker id="arrow" viewBox="0 0 12 10" refX="10" refY="5"
          markerWidth="8" markerHeight="6" orient="auto-start-reverse"
          fill="none" stroke="#eee0cc" stroke-width="1.5"
          stroke-linecap="round" stroke-linejoin="round">
    <path d="M 1,1 L 10,5 L 1,9"/>
  </marker>
</defs>
<path d="M 336,134 C 338,150 334,162 336,176"
      stroke="#eee0cc" stroke-width="2" fill="none"
      stroke-linecap="round" marker-end="url(#arrow)"/>
```

**Rotation**: Apply slight rotation transforms (0.3-1.2 degrees, alternating positive/negative) to element groups.

```xml
<g transform="rotate(-0.5, 336, 630)">
  
</g>
```

**Filters**: Use roughen filter for organic texture, glow filter for accent elements.

```xml
<defs>
  <filter id="roughen" x="-2%" y="-2%" width="104%" height="104%">
    <feTurbulence type="turbulence" baseFrequency="0.03" numOctaves="2" result="noise" seed="5"/>
    <feDisplacementMap in="SourceGraphic" in2="noise" scale="1.2" xChannelSelector="R" yChannelSelector="G"/>
  </filter>
  <filter id="glow">
    <feGaussianBlur stdDeviation="3" result="blur"/>
    <feMerge><feMergeNode in="blur"/><feMergeNode in="SourceGraphic"/></feMerge>
  </filter>
</defs>
```

**Dashed connectors**: Use `stroke-dasharray="4 4"` or `stroke-dasharray="8 5"` for secondary/feedback paths.

**Underlines**: Sketchy path under headings, same color as heading, reduced opacity.

```xml
<path d="M 172,46 C 240,48 360,44 500,47"
      stroke="#e88072" stroke-width="1.5" fill="none"
      stroke-linecap="round" opacity="0.6"/>
```

### ViewBox Sizing

All SVGs use `viewBox` with no fixed `width`/`height` (responsive). Standard slide SVG viewBox: `0 0 672 400` (fits 16:9 slides). Adjust height as needed but keep width at 672 for consistency.

### SVG Checklist (per SVG)

- [ ] No fixed `width`/`height` on root `<svg>` -- only `viewBox`
- [ ] No clean `<rect>` for content boxes -- use sketchy `<path>`
- [ ] No clean `<line>` for visible strokes -- use sketchy `<path>`
- [ ] Uses `Caveat` font family with fallback stack
- [ ] Slight rotation on element groups (varied, not uniform)
- [ ] Colors from the dark palette only (`#eee0cc`, `#e88072`, `#4ade80`, `#f87171`, `#888888`)
- [ ] No background rect (transparent -- sits on dark slide background)
- [ ] Arrows use marker-end or inline chevron paths
- [ ] Roughen filter on organic elements
- [ ] Self-contained: no external dependencies

---

## sli.dev Theme Configuration

Every deck starts with this frontmatter block:

```md
---
theme: default
title: "DECK TITLE"
info: |
  DECK SUBTITLE OR DESCRIPTION
class: text-white
drawings:
  persist: false
transition: fade-out
mdc: true
---
```

### Global Style Block

Place this on the first slide (cover slide), after the frontmatter:

```html
<style>
@import url('https://fonts.googleapis.com/css2?family=Caveat:wght@400;700&display=swap');

:root {
  --slidev-theme-primary: #e88072;
  --slidev-theme-background: #0d1117;
}

.slidev-layout {
  background: #0d1117 !important;
  color: #e6edf3;
}

h1, h2, h3 {
  font-family: 'Caveat', 'Comic Sans MS', cursive !important;
  color: #eee0cc !important;
}

h1 { font-size: 2.5rem !important; }

p, li, td {
  color: #e6edf3;
  font-size: 1.1rem;
  line-height: 1.6;
}

strong { color: #e88072; }

.coral { color: #e88072; }
.green { color: #4ade80; }
.red { color: #f87171; }
.muted { color: #888888; }

blockquote {
  border-left: 3px solid #e88072;
  padding-left: 1rem;
  color: #888888;
  font-style: italic;
}

code {
  background: rgba(232,128,114,0.1);
  color: #e88072;
  padding: 0.1em 0.3em;
  border-radius: 3px;
}

table {
  width: 100%;
}

th {
  color: #e88072;
  border-bottom: 2px solid rgba(232,128,114,0.3);
}

td {
  border-bottom: 1px solid rgba(238,224,204,0.1);
}

.svg-container {
  display: flex;
  justify-content: center;
  align-items: center;
  width: 100%;
  max-width: 672px;
  margin: 0 auto;
}

.svg-container svg {
  width: 100%;
  height: auto;
}

.stat-big {
  font-family: 'Caveat', cursive;
  font-size: 4rem;
  color: #e88072;
  font-weight: bold;
  line-height: 1;
}

.stat-label {
  color: #888888;
  font-size: 0.9rem;
}
</style>
```

---

## Deck Types

### 1. Product Pitch Deck

**Audience**: Warm leads, partners, potential customers, investors.

**Story arc**: Problem --> Evidence --> Solution --> How It Works --> Proof --> Pricing --> CTA.

**Length**: 14-18 slides.

**Pipeline data sources**:
- `signal-scan` output: pain signals, demand signals, evidence quotes
- `decision-log` output: SDP scores, thesis statement
- `persona-extract` output: pain stories, decision triggers, persona names
- `offer-scope` output: pricing, value equation, distribution plan
- `pitch` output: landing page copy, launch posts, positioning

#### Slide Sequence

**Slide 1: Cover**
```md
---
layout: cover
background: '#0d1117'
---

# PRODUCT NAME

TAGLINE FROM PITCH POSITIONING COPY

**Slide 2: The Problem**
```md
---
layout: center
---
```
- **SVG**: "Pain landscape" -- 3-4 sketchy boxes showing the top pain points from signal-scan PAIN signals. Each box has a short pain quote from real evidence. Boxes connected by dashed lines to a central "problem" label. Red (`#f87171`) accent on the most acute pain.
- **Text**: One headline sentence naming the problem.
- **Data source**: `signal-scan` PAIN signals (top 3-4 by intensity).
- Speaker notes: read the actual pain quotes, name the Reddit threads or communities where these came from.

**Slide 3: The Evidence**
```md
---
layout: center
---
```
- **SVG**: "Signal dashboard" -- A sketchy 2x3 or 3x2 grid of stat cards. Each card shows one metric: number with a label. Examples: "47 upvotes on 'I spent 4 hours...'", "$19-49 Udemy courses, 10K+ students", "3 Reddit threads same complaint". Use numbered callouts (coral circles).
- **Text**: One headline: "We didn't guess. We measured."
- **Data source**: `signal-scan` evidence quotes and metrics across all signal types.
- Speaker notes: walk through each stat, explain why this is not a guess but a measured signal.

**Slide 4: Who Feels This**
```md
---
layout: two-cols
---
```
- **SVG (left column)**: "Persona sketch" -- A hand-drawn persona card with name, role title, and 2-3 bullet pain points. Use a sketchy person icon (circle head + bezier body outline) at the top.
- **Text (right column)**: Key persona details: role, pain story snippet (in their words), decision trigger, what they've tried that failed.
- **Data source**: `persona-extract` primary persona.
- Speaker notes: tell the persona's pain story. Use their actual language.

**Slide 5: Why Now**
```md
---
layout: center
---
```
- **SVG**: "Timing diagram" -- A sketchy timeline showing 3-4 events that create the current window of opportunity. Use the numbered-circle style from the pipeline diagram. Connect events with bezier arrows. The final event (now) gets a glow filter.
- **Text**: One sentence: why this problem is acute right now and not two years ago.
- **Data source**: `signal-scan` meta-signal synthesis + `decision-log` thesis.
- Speaker notes: explain the structural shift that created this opportunity.

**Slide 6: The Solution**
```md
---
layout: center
---
```
- **SVG**: "Solution overview" -- Central product name in a glowing sketchy box, surrounded by 3-4 feature/benefit boxes connected by arrows. Each benefit box uses green (`#4ade80`) accent to contrast with the red pain boxes from slide 2.
- **Text**: One headline naming the product. No bullet lists.
- **Data source**: `offer-scope` product description and value proposition.
- Speaker notes: describe the product in one sentence, then explain each benefit area.

**Slide 7: How It Works (1 of 2)**
```md
---
layout: center
---
```
- **SVG**: "Process flow" -- A 3-5 step horizontal or vertical flow showing how the product works. Each step is a numbered coral circle with a label. Steps connected by bezier arrows. Keep it high-level (user perspective, not technical internals).
- **Text**: None or minimal -- the SVG IS the content.
- **Data source**: `offer-scope` curriculum/product structure.
- Speaker notes: walk through each step. Explain what the user does and what they get.

**Slide 8: How It Works (2 of 2) -- Optional**
```md
---
layout: two-cols
---
```
- **SVG (left)**: "Before/After" comparison -- Two-column sketchy comparison. Left side (red accents): current state pain. Right side (green accents): after-state with product. Use the abstraction-fork SVG style from the portfolio.
- **Text (right)**: 3 bullet contrasts: "Before: X. After: Y."
- **Data source**: `persona-extract` pain stories vs. `offer-scope` outcomes.
- Speaker notes: contrast the current workflow with the product workflow.

**Slide 9: Social Proof / Validation**
```md
---
layout: center
---
```
- **SVG**: "Evidence wall" -- 3-4 quote boxes arranged in a staggered layout. Each box contains a real quote from signal-scan evidence (Reddit, HN, etc.) with source attribution. Boxes have slight rotation for the napkin-on-desk feel.
- **Text**: Headline: "What they're already saying."
- **Data source**: `signal-scan` raw signal evidence quotes.
- Speaker notes: these are real quotes from real people. Name the sources.

**Slide 10: Competitive Landscape**
```md
---
layout: center
---
```
- **SVG**: "Competitive map" -- 2x2 quadrant diagram (sketchy axes, labeled). Plot competitors as labeled dots. Product occupies the desirable quadrant (upper-right or equivalent). Axes labels come from the two most important differentiation dimensions from `signal-scan` COMPETITIVE signals.
- **Text**: One sentence positioning statement.
- **Data source**: `signal-scan` COMPETITIVE signals + `decision-log` competition gap analysis.
- Speaker notes: explain why competitors miss the mark. Reference specific gaps.

**Slide 11: Revenue Model**
```md
---
layout: center
---
```
- **SVG**: "Revenue funnel" -- A sketchy funnel or stacked bar showing the revenue path: audience size --> conversion rate --> revenue per customer --> total addressable. Use green for revenue numbers. Use coral numbered circles for each stage.
- **Text**: Key number: projected revenue or price point.
- **Data source**: `offer-scope` revenue model and pricing.
- Speaker notes: walk through the math. Ground every number in evidence from the pipeline.

**Slide 12: Pricing**
```md
---
layout: center
---
```
- **SVG**: "Pricing card" -- A sketchy pricing card with product name, price, and 4-5 included items (checkmarks in green). If there are tiers, show 2-3 side-by-side cards with the recommended tier highlighted (glow filter).
- **Text**: Price and what's included. Nothing else.
- **Data source**: `offer-scope` pricing and product inclusions.
- Speaker notes: justify the price by referencing competitive pricing from SPEND signals.

**Slide 13: Risk Mitigation**
```md
---
layout: center
---
```
- **SVG**: "Kill criteria dashboard" -- 3-4 sketchy boxes, each containing one kill criterion with its metric and deadline. Use red (`#f87171`) for the thresholds, muted text for the descriptions. This shows the audience you have exit discipline.
- **Text**: Headline: "We know when to stop."
- **Data source**: `decision-log` kill criteria + `offer-scope` kill criteria.
- Speaker notes: explain each criterion. This builds trust -- you're not blindly optimistic.

**Slide 14: CTA**
```md
---
layout: center
---
```
- **SVG**: Minimal -- just a large hand-drawn arrow pointing right, with glow filter. The arrow is the visual.
- **Text**: Clear call to action. One sentence. URL or next step. Large font.
- **Data source**: `pitch` CTA copy.
- Speaker notes: tell them exactly what to do next. One action, no ambiguity.

**Slide 15: Appendix / Q&A -- Optional**
```md
---
layout: center
---

# Q&A

```
- No SVG. Clean slide.
- Speaker notes: reference specific pipeline artifacts for anticipated questions.

---

### 2. Validation Summary Deck

**Audience**: Internal team, co-founders, advisors.

**Story arc**: Opportunity --> Signals --> Decision --> Personas --> Offer --> Go/No-Go.

**Length**: 12-16 slides.

**Pipeline data sources**: All pipeline outputs for the product being reviewed.

#### Slide Sequence

**Slide 1: Cover**
```md
---
layout: cover
background: '#0d1117'
---

# Validation Summary: PRODUCT NAME

Pipeline review -- DATE

**Slide 2: Pipeline Overview**
```md
---
layout: center
---
```
- **SVG**: "Pipeline stages" -- The 6-stage pipeline diagram (Signal Scan --> Decision Log --> Persona Extract --> SWOT --> Offer Scope --> Pitch). Numbered coral circles, sketchy boxes, bezier arrows. Highlight which stages are complete (green checkmark or green accent) vs. pending (muted). Use the exact style from the portfolio's `pipeline-architecture.svg`.
- **Text**: None.
- **Data source**: Check which pipeline artifacts exist in the vault for this product.
- Speaker notes: walk through the pipeline. Explain what each stage produces.

**Slide 3: Signal Scan Results**
```md
---
layout: center
---
```
- **SVG**: "Signal heatmap" -- A 7-row table-like visualization showing the 7 signal types (PAIN, DEMAND, SPEND, COMPETITIVE, SENTIMENT, BEHAVIOR, AUDIENCE) with a strength indicator for each (1-5 filled circles or bars). Use coral for strong signals, muted for weak.
- **Text**: Domain name and scan date.
- **Data source**: `signal-scan` signal counts and intensities.
- Speaker notes: which signals are strong, which are weak, what gaps exist.

**Slide 4: Top Signals (Detail)**
```md
---
layout: center
---
```
- **SVG**: "Evidence cards" -- 3-5 signal evidence cards in a staggered layout. Each card has: signal type label, evidence quote, source, intensity score. Most intense signal gets glow treatment.
- **Text**: None.
- **Data source**: `signal-scan` top 3-5 signals by intensity.
- Speaker notes: read each quote. Explain why it matters.

**Slide 5: Opportunity Ranking**
```md
---
layout: center
---
```
- **SVG**: "Ranked opportunities" -- A horizontal bar chart (sketchy bars) showing 3-5 opportunities ranked by score. Use coral fill proportional to score. The chosen opportunity gets a green border.
- **Text**: Table below SVG with opportunity names and scores.
- **Data source**: `signal-scan` opportunity ranking.
- Speaker notes: explain the scoring dimensions. Why the top opportunity is top.

**Slide 6: Decision Frame**
```md
---
layout: center
---
```
- **SVG**: "Decision snapshot" -- A sketchy card showing: Decision statement, time horizon, success metric, constraints. Clean layout, one item per row. Use the two-column comparison style if comparing go vs. no-go.
- **Text**: The decision statement in large text.
- **Data source**: `decision-log` Phase 1 (decision frame).
- Speaker notes: why this frame, what alternatives were considered.

**Slide 7: SDP Scoring**
```md
---
layout: center
---
```
- **SVG**: "Radar chart" -- A sketchy radar/spider chart showing 6 SDP dimensions (Pain, Spend, Edge, Speed, Gap, Audience) for the chosen opportunity. Plot as a filled polygon with coral accent. Optionally overlay the runner-up in muted gray.
- **Text**: SDP score as a large stat: "42/60".
- **Data source**: `decision-log` Phase 3 (SDP scores).
- Speaker notes: walk through each dimension. Cite the evidence for each score.

**Slide 8: Bias Check + Pre-Mortem**
```md
---
layout: two-cols
---
```
- **SVG (left)**: "Bias check" -- A sketchy checklist with 4 items (confirmation, availability, anchoring, excitement). Each has a green check or red flag. Simple layout.
- **Text (right)**: Pre-mortem scenarios: 3 numbered failure scenarios.
- **Data source**: `decision-log` Phases 5-6.
- Speaker notes: be honest about the biases. Which flags were raised and why we proceeded anyway.

**Slide 9: Persona Deep-Dive**
```md
---
layout: center
---
```
- **SVG**: "Persona profile" -- Hand-drawn persona card with: name, role, a simplified Four Forces diagram (push/pull/anxiety/habit as four quadrants around a central label). Use green for push/pull (driving change), red for anxiety/habit (resisting change).
- **Text**: None -- SVG carries all content.
- **Data source**: `persona-extract` primary persona + Four Forces analysis.
- Speaker notes: tell the persona's story. Use their actual words from pain stories.

**Slide 10: Offer Spec**
```md
---
layout: center
---
```
- **SVG**: "Offer blueprint" -- A sketchy product card showing: product name, format, price, ship time, distribution channel. Arranged as a labeled box with sections. Hormozi value equation scores shown as a small sub-diagram (dream outcome, perceived likelihood, time delay, effort/sacrifice).
- **Text**: None.
- **Data source**: `offer-scope` product spec and value equation.
- Speaker notes: explain the value equation. Why this format, this price, this channel.

**Slide 11: Kill Criteria**
```md
---
layout: center
---
```
- **SVG**: "Tripwire board" -- 3-4 sketchy card strips, each showing one kill criterion with metric, threshold, and deadline. Use red accent for thresholds. Visual metaphor: tripwires that trigger an exit.
- **Text**: Headline: "When we walk away."
- **Data source**: `decision-log` + `offer-scope` kill criteria.
- Speaker notes: commit to these out loud. This is a contract with the team.

**Slide 12: Go/No-Go Recommendation**
```md
---
layout: center
---
```
- **SVG**: A single large word -- "GO" in green with glow filter, or "NO-GO" in red with glow filter, or "CONDITIONAL" in coral. Below it, 2-3 condition bullets if conditional.
- **Text**: The recommendation, one sentence.
- **Data source**: Synthesized from all pipeline outputs.
- Speaker notes: summarize the case. Restate the key evidence. Ask for the decision.

---

### 3. Technical Architecture Deck

**Audience**: Engineering teams, technical reviewers, architecture review boards.

**Story arc**: Problem --> Architecture --> Key Decisions --> Trade-offs --> Results.

**Length**: 12-16 slides.

**Pipeline data sources**: Primarily system knowledge + any relevant pipeline outputs for product context. This deck type may be used independently of the hunter pipeline.

#### Slide Sequence

**Slide 1: Cover**
```md
---
layout: cover
background: '#0d1117'
---

# SYSTEM NAME: Architecture Overview

DATE

**Slide 2: The Problem**
```md
---
layout: center
---
```
- **SVG**: "Problem diagram" -- Show the system constraint or scaling challenge. Example: "stdio at scale" diagram showing 1 consumer (fine), 3 consumers (not fine), N consumers (dead). Use the portfolio's scaling diagram style.
- **Text**: One sentence naming the engineering challenge.
- Speaker notes: technical context for why this architecture exists.

**Slide 3: Architecture Overview**
```md
---
layout: center
---
```
- **SVG**: "System architecture" -- Full system diagram with layers, services, data stores, external inputs/outputs. Use the pipeline-architecture.svg style: numbered coral circles for layers, sketchy boxes, bezier arrows for data flow, dashed lines for external boundaries. This is the hero SVG of the deck -- spend the most detail here.
- **Text**: None.
- Speaker notes: walk through each layer top to bottom. Name the technologies at each layer.

**Slide 4: Data Flow**
```md
---
layout: center
---
```
- **SVG**: "Data flow diagram" -- Show how data moves through the system. Use animated-dash bezier paths (via `stroke-dasharray` + descriptions in speaker notes since slidev can add CSS animations). Label each flow with what data travels on it.
- **Text**: None.
- Speaker notes: trace a single request through the system end to end.

**Slide 5: Key Decision 1**
```md
---
layout: two-cols
---
```
- **SVG (left)**: "Decision fork" -- The abstraction-fork style: two columns (Option A vs. Option B), with trade-offs listed. Chosen option in green, rejected in red/muted.
- **Text (right)**: Decision rationale: why Option A won. 2-3 bullet reasons.
- Speaker notes: what would have happened if we'd chosen the other option.

**Slide 6: Key Decision 2**
```md
---
layout: two-cols
---
```
- Same structure as Slide 5 for the second key decision.
- Speaker notes: pattern of decisions tells a story about engineering values.

**Slide 7: Key Decision 3 -- Optional**
```md
---
layout: two-cols
---
```
- Same structure. Use only if there's a third critical decision worth presenting.

**Slide 8: The Abstraction That Paid Off**
```md
---
layout: center
---
```
- **SVG**: "Import swap" -- Show the before/after of the key abstraction. The portfolio's import-swap.svg style: two code-like boxes showing the one-line change that enabled the architecture shift. Green accent on the "after".
- **Text**: One sentence explaining the leverage the abstraction provides.
- Speaker notes: explain why this abstraction was worth the investment. What it unlocks.

**Slide 9: Trade-offs**
```md
---
layout: center
---
```
- **SVG**: "Trade-off matrix" -- A sketchy 2x2 or 3-column comparison: What we gained vs. What it cost vs. What we'd do differently. Use green for gains, red for costs, coral for improvements.
- **Text**: None.
- Speaker notes: be honest about the costs. Engineering credibility comes from acknowledging trade-offs.

**Slide 10: Performance / Results**
```md
---
layout: center
---
```
- **SVG**: "Metrics dashboard" -- 3-4 large stat boxes with key performance metrics. Example: "< 50ms p99 latency", "3x throughput", "1 import swap to migrate". Use large numbers in coral, labels in muted.
- **Text**: None.
- Speaker notes: explain how each metric was measured. What the baseline was.

**Slide 11: What's Next**
```md
---
layout: center
---
```
- **SVG**: "Roadmap sketch" -- A sketchy timeline with 3-4 upcoming milestones. Current position marked with a glowing dot. Future items in muted/dashed style.
- **Text**: None.
- Speaker notes: explain priorities and sequencing.

**Slide 12: Q&A**
```md
---
layout: center
---

# Questions?

```

---

### 4. Strategy Deck

**Audience**: Business stakeholders, advisors, internal leadership.

**Story arc**: Market --> Positioning --> Revenue --> Execution Plan --> Kill Criteria.

**Length**: 14-18 slides.

**Pipeline data sources**:
- `signal-scan`: market signals, competitive landscape, demand evidence
- `decision-log`: strategic decisions, SDP scores
- `persona-extract`: target customer profiles
- `offer-scope`: revenue model, pricing, distribution
- `pitch`: positioning copy, launch plan

#### Slide Sequence

**Slide 1: Cover**
```md
---
layout: cover
background: '#0d1117'
---

# PRODUCT NAME: Strategy

QUARTER/YEAR

**Slide 2: Market Thesis**
```md
---
layout: center
---
```
- **SVG**: "Market shift" -- A before/after timeline showing the structural change that creates the opportunity. Left side (muted, past): how the market used to work. Right side (coral, present): what changed. Bezier arrow bridging them with the thesis statement as a label.
- **Text**: The meta-signal from signal-scan in one sentence.
- **Data source**: `signal-scan` meta-signal synthesis.
- Speaker notes: explain the structural failure that creates this opening.

**Slide 3: Market Size**
```md
---
layout: center
---
```
- **SVG**: "TAM/SAM/SOM" -- Three concentric sketchy circles (largest to smallest). Labels: Total Addressable Market (outer), Serviceable Addressable Market (middle), Serviceable Obtainable Market (inner). Numbers inside each.
- **Text**: The SOM number as a large stat.
- **Data source**: `signal-scan` DEMAND + SPEND signals for market sizing.
- Speaker notes: explain how each tier was calculated. Ground in evidence.

**Slide 4: Competitive Landscape**
```md
---
layout: center
---
```
- **SVG**: "Competitive map" -- Same 2x2 quadrant as product pitch deck slide 10, but with more detail: competitor names, their weaknesses annotated with callout lines.
- **Text**: Positioning statement.
- **Data source**: `signal-scan` COMPETITIVE signals + `decision-log` competition gap.
- Speaker notes: name each competitor's weakness. Explain why they can't or won't fix it.

**Slide 5: Target Customer**
```md
---
layout: center
---
```
- **SVG**: "Customer profile" -- Primary persona sketch with pain story headline, role, budget, decision-making authority. Simplified version of the persona card.
- **Text**: Who we sell to, in one sentence.
- **Data source**: `persona-extract` primary persona.
- Speaker notes: tell the pain story. Make the audience feel the problem.

**Slide 6: Positioning**
```md
---
layout: center
---
```
- **SVG**: "Positioning statement" -- A hand-drawn card with the positioning framework: "For [target], who [pain], [product] is a [category] that [key benefit]. Unlike [competitors], we [differentiator]." Each bracketed section in coral, the connective words in muted.
- **Text**: None -- the SVG IS the positioning statement.
- **Data source**: `pitch` positioning copy + `offer-scope` value proposition.
- Speaker notes: explain each element of the positioning. Why these specific words.

**Slide 7: Revenue Model**
```md
---
layout: center
---
```
- **SVG**: "Revenue architecture" -- A sketchy flow showing: pricing tiers (if any), revenue per customer, expected volume, total revenue. Use stacked bars or a funnel. Green for revenue numbers.
- **Text**: Key revenue metric as a large stat.
- **Data source**: `offer-scope` revenue model.
- Speaker notes: walk through the math. Explain assumptions.

**Slide 8: Distribution Strategy**
```md
---
layout: center
---
```
- **SVG**: "Distribution channels" -- 3-4 channel boxes arranged around a central product box. Each channel box labeled with the platform (Reddit, LinkedIn, email, etc.) and the strategy (value-first content, launch post, etc.). Arrows from channels to product.
- **Text**: None.
- **Data source**: `offer-scope` distribution plan + `pitch` launch posts.
- Speaker notes: explain each channel. Which is primary, which is secondary.

**Slide 9: Execution Calendar**
```md
---
layout: center
---
```
- **SVG**: "Sprint timeline" -- A sketchy Gantt-like timeline showing build phases, launch date, and post-launch milestones. Use coral for current phase, green for completed, muted for future. Include key dates.
- **Text**: Launch date as a large stat.
- **Data source**: `offer-scope` execution calendar + `pitch` launch checklist.
- Speaker notes: walk through each phase. What's the critical path.

**Slide 10: Pricing Strategy**
```md
---
layout: center
---
```
- **SVG**: "Pricing comparison" -- Side-by-side sketchy cards: our pricing vs. competitor pricing vs. what the market currently pays. Green accent on ours to show value positioning.
- **Text**: Price point and justification.
- **Data source**: `offer-scope` pricing + `signal-scan` SPEND signals.
- Speaker notes: justify the price by referencing SPEND evidence.

**Slide 11: Risk Analysis**
```md
---
layout: center
---
```
- **SVG**: "Risk matrix" -- A sketchy 2x2 grid (likelihood vs. impact). Plot 3-5 risks as labeled dots. High-likelihood + high-impact quadrant in red. Include mitigation labels as annotation callouts.
- **Text**: None.
- **Data source**: `decision-log` pre-mortem scenarios.
- Speaker notes: explain each risk and its mitigation. Be direct about what could go wrong.

**Slide 12: Kill Criteria**
```md
---
layout: center
---
```
- **SVG**: Same "Tripwire board" from validation summary deck.
- **Text**: Headline: "Exit discipline."
- **Data source**: `decision-log` + `offer-scope` kill criteria.
- Speaker notes: commit to these publicly.

**Slide 13: Success Metrics**
```md
---
layout: center
---
```
- **SVG**: "Scorecard" -- 4-5 metric boxes with: metric name, target value, measurement method, timeline. Green accent on targets.
- **Text**: None.
- **Data source**: `decision-log` success metrics + `offer-scope` kill criteria (inverted).
- Speaker notes: how and when each metric will be measured.

**Slide 14: Ask / CTA**
```md
---
layout: center
---
```
- **SVG**: Minimal -- large hand-drawn arrow or a single word.
- **Text**: What you need from the audience: decision, resources, feedback, approval.
- Speaker notes: be specific. What exactly do you need and by when.

---

## Workflow

```
User request (deck type + product + audience)
    |
Phase 1: Gather pipeline data (glob vault, read artifacts)
    |
Phase 2: Map data to slide sequence (fill each slide's data fields)
    |
Phase 3: Generate SVGs (use sketches skill style, inline in markdown)
    |
Phase 4: Assemble slides.md (frontmatter + global styles + slides)
    |
Phase 5: Add speaker notes (every slide)
    |
Phase 6: Quality check (run checklist)
    |
Phase 7: Write Obsidian summary note
    |
Output: slides.md + Obsidian note
```

### Phase 1: Gather Pipeline Data

Glob the vault for all pipeline artifacts related to the product:

```
Glob: {vault}/Admin/Product-Discovery/**/{product-slug}*.md
Glob: {vault}/Admin/Product-Discovery/**/{product-slug}*.json
```

Read each artifact and extract the fields needed for the chosen deck type. If an artifact is missing, note which slides will have incomplete data and ask the user whether to proceed with what's available or wait.

### Phase 2: Map Data to Slides

For each slide in the chosen deck type's sequence, map specific pipeline data to the slide's content fields:

```
Slide N:
  - headline: [from which artifact, which field]
  - SVG description: [what to draw, what data to embed]
  - text content: [bullets, stats, quotes -- with source]
  - speaker notes: [context, evidence, talking points]
```

If a data point is not available from the pipeline, DO NOT fabricate it. Either skip the slide or mark it as "[DATA NEEDED: describe what's missing]".

### Phase 3: Generate SVGs

For each slide that needs an SVG, generate it inline following the SVG Style Contract above. Every SVG must:

1. Use the dark-theme palette (no light backgrounds)
2. Use sketchy primitives (no clean rects or lines)
3. Use Caveat font family
4. Have a `viewBox` of `0 0 672 NNN` (width 672, height varies)
5. Be self-contained (no external references)
6. Have NO background rect (transparent, sits on `#0d1117` slide background)

For complex SVGs (architecture diagrams, pipeline flows, competitive maps), use the `sketches` skill's technique reference:

```
Read /Users/peleke/Documents/Projects/skills/skills/custom/sketches/references/svg-patterns.md
```

Adapt the whiteboard palette patterns to the dark palette colors.

### Phase 4: Assemble slides.md

Combine all slides into a single `slides.md` file:

```md
---
theme: default
title: "DECK TITLE"
info: |
  DECK DESCRIPTION
class: text-white
drawings:
  persist: false
transition: fade-out
mdc: true
---

# Slide 1 Title

content

---

# Slide 2 Title

content

---
layout: two-cols
---

# Slide 3 Left

::right::

Right content

---

```

Slides are separated by `---` on their own line. Layout directives go in the slide's frontmatter block (the `---` block at the top of each slide).

### Phase 5: Add Speaker Notes

Every slide MUST have speaker notes. Add them as HTML comments at the end of each slide:

```md
# Slide Title

content

```

Speaker notes should:
- Provide context the slide doesn't show
- Include exact data references (which pipeline artifact, which field)
- Suggest transitions to the next slide
- Anticipate audience questions
- Never exceed 6 bullet points

### Phase 6: Quality Check

Run the full checklist (see Quality Checklist below) before delivering.

### Phase 7: Write Obsidian Summary Note

After generating the deck, write a summary note to the vault.

---

## Pipeline Integration

### Data Extraction Map

This table shows which pipeline output fields feed which deck elements:

| Pipeline Artifact | Key Fields to Extract | Used In (Deck Slides) |
|---|---|---|
| `signal-scan` | `signals[type=PAIN]` (quotes, intensity) | Problem, Evidence, Social Proof |
| `signal-scan` | `signals[type=SPEND]` (prices, platforms) | Revenue, Pricing, Market Size |
| `signal-scan` | `signals[type=COMPETITIVE]` (competitors, gaps) | Competitive Landscape |
| `signal-scan` | `signals[type=DEMAND]` (volume, trends) | Market Size, Evidence |
| `signal-scan` | `meta_signal` | Market Thesis, Why Now |
| `signal-scan` | `opportunities[]` (ranked, scored) | Opportunity Ranking |
| `decision-log` | `decision_frame` (question, horizon, metric) | Decision Frame |
| `decision-log` | `sdp_scores` (6 dimensions, justifications) | SDP Scoring |
| `decision-log` | `bias_check` (pass/flag, notes) | Bias Check |
| `decision-log` | `pre_mortem` (failure scenarios) | Risk Analysis, Pre-Mortem |
| `decision-log` | `kill_criteria` (measurable, time-bounded) | Kill Criteria |
| `decision-log` | `thesis` | Market Thesis |
| `persona-extract` | `personas[].pain_stories` | Who Feels This, Problem |
| `persona-extract` | `personas[].four_forces` | Persona Deep-Dive |
| `persona-extract` | `personas[].decision_triggers` | Who Feels This |
| `persona-extract` | `personas[].profile` (role, context) | Target Customer |
| `offer-scope` | `product` (name, format, price) | Solution, Pricing |
| `offer-scope` | `value_equation` (Hormozi scores) | Offer Spec |
| `offer-scope` | `revenue_model` | Revenue Model |
| `offer-scope` | `distribution_plan` | Distribution Strategy |
| `offer-scope` | `execution_calendar` | Execution Calendar |
| `offer-scope` | `kill_criteria` | Kill Criteria |
| `pitch` | `landing_page_copy` | Positioning, CTA |
| `pitch` | `launch_posts` | Distribution |
| `pitch` | `positioning` | Positioning Statement |
| `pitch` | `email_sequence` | (appendix if needed) |

### Missing Data Protocol

If a pipeline artifact is missing:

1. Check if the user has the raw data available elsewhere
2. If not, mark affected slides with ``
3. Generate the deck with placeholder SVGs showing "Data pending: run [skill name]" in muted text
4. List all missing artifacts in the Obsidian summary note

---

## Output

### 1. slides.md

**Location**: Written to the user's specified directory, or defaulting to:
```
{project-root}/decks/{product-slug}-{deck-type}/slides.md
```

This file is ready to run with:
```bash
npx slidev slides.md
```

No `package.json` is required -- `npx slidev` handles dependency resolution. If the user wants a persistent setup, suggest:

```bash
npm init slidev@latest
# then replace slides.md with the generated file
```

### 2. Obsidian Summary Note

**Vault path**: `{vault}/Admin/Product-Discovery/Decks/{product-slug}-{deck-type}-{YYYY-MM-DD}.md`

Create the `Decks` folder if it doesn't exist:
```bash
mkdir -p "{vault}/Admin/Product-Discovery/Decks"
```

**Frontmatter**:
```yaml
---
type: deck
date: YYYY-MM-DD
status: draft
deck_type: product-pitch | validation-summary | technical-architecture | strategy
product: PRODUCT NAME
audience: TARGET AUDIENCE
tags:
  - hunter/deck
  - hunter/domain/{domain-slug}
  - hunter/opportunity/{opportunity-slug}
signal_scan_ref: "{signal-scan-slug}"
decision_ref: "{decision-slug}"
persona_ref: "{persona-slug}"
offer_ref: "{offer-slug}"
pitch_ref: "{pitch-slug}"
---
```

**Body structure**:
```markdown
# Deck: PRODUCT NAME -- DECK TYPE

**Generated**: YYYY-MM-DD
**Audience**: TARGET AUDIENCE
**Slides**: N slides
**Key message**: THE ONE SENTENCE THE AUDIENCE SHOULD REMEMBER

## Slide Outline

1. **Cover**: PRODUCT NAME
2. **The Problem**: [headline]
3. **The Evidence**: [headline]
...

## Pipeline Data Used

- Signal Scan: [[signal-scan-ref]]
- Decision Log: [[decision-ref]]
- Persona Extract: [[persona-ref]]
- Offer Scope: [[offer-ref]]
- Pitch: [[pitch-ref]]

## Missing Data

- [list any pipeline artifacts that were not available]

## File Location

`{absolute-path-to-slides.md}`

## Notes

[Any generation notes, caveats, or follow-up items]
```

---

## Quality Checklist

Run before delivering any deck:

### Content Quality
- [ ] Every slide has a single clear message (no slide tries to say two things)
- [ ] No slide has more than 6 lines of text (SVGs carry the story)
- [ ] Story arc is complete: problem --> evidence --> solution --> proof --> action
- [ ] All data traces to pipeline outputs (no fabricated stats)
- [ ] Fabricated data is flagged with `[DATA NEEDED]` markers
- [ ] Speaker notes exist on every slide
- [ ] Speaker notes reference specific pipeline artifacts

### SVG Quality
- [ ] Every key slide (problem, evidence, solution, architecture, pricing, CTA) has an inline SVG
- [ ] Total SVG count: 5-10 per deck (not fewer, not more)
- [ ] SVGs follow the dark-theme sketchy style (no clean primitives)
- [ ] No SVG uses light backgrounds (all transparent for dark slides)
- [ ] All SVGs use `viewBox` with no fixed dimensions
- [ ] All SVGs use Caveat font family
- [ ] Numbered callouts use coral fill circles with white text
- [ ] Slight rotation applied to element groups
- [ ] Colors restricted to the dark palette

### sli.dev Technical
- [ ] Frontmatter is valid YAML
- [ ] Slide separators are `---` on their own line
- [ ] Layout directives are in correct position (slide frontmatter block)
- [ ] Global style block appears only on the first slide
- [ ] `transition: fade-out` is set
- [ ] `class: text-white` is set on root frontmatter
- [ ] File is runnable with `npx slidev slides.md` without errors

### Deck Structure
- [ ] Total deck length: 12-20 slides (not more)
- [ ] Cover slide is clean (no SVG, just title + tagline)
- [ ] CTA slide is clear and actionable (one action, no ambiguity)
- [ ] Dark theme consistent throughout (no white backgrounds leak through)
- [ ] Slide transitions feel natural (each slide's speaker notes suggest a transition)

### Obsidian Output
- [ ] Summary note written to `{vault}/Admin/Product-Discovery/Decks/`
- [ ] Frontmatter includes all required fields (type, date, status, tags, refs)
- [ ] Slide outline in summary note matches actual deck
- [ ] Missing data documented
- [ ] File location path is absolute