storyline
Designs lecture storyline and slide deck composition. Given research_notes.md (or direct topic input), produces storyline.md and slide_deck.md. Responds to "make storyline", "compose slide deck", "structure chapters" (Korean: "make storyline", "compose slide deck", "restructure chapters"). Stages 2-3 of the lecture workflow.
- License 许可
- AGPL-3.0
- 最近更新
- 2026-11-14
- Stars
- ★ 25
- Forks
- 9
01 前言 Front Matter
- name
- storyline
- owner
- storyline
- scope
- co-deck
- inputs
- [ "research-notes" ]
- status
- active
- outputs
- [ "storyline-outline", "slide-deck-draft" ]
- version
- 1.2.0
- relates_to
- [ { "type": "follows", "skill": "design" } ]
- description
- Designs lecture storyline and slide deck composition. Given research_notes.md (or direct topic input), produces storyline.md and slide_deck.md. Responds to "make storyline", "compose slide deck", "structure chapters" (Korean: "make storyline", "compose slide deck", "restructure chapters"). Stages 2-3 of the lecture workflow.
- last_reviewed
- 1781827200000
- prerequisites
- research
03 SKILL.md 正文
Context
Designs the lecture narrative flow (storyline.md) and per-slide content (slide_deck.md). slide_deck.md is the direct input consumed by the Build Agent for HTML generation. Invoked at Stages 2-3, after research and optional source verification.
When to Use
- PM Agent dispatches after Stage 1 (research)
- User says "make storyline"
- User says "compose slide deck"
- User says "restructure chapters"
Execution Steps
Step 1: Confirm Inputs
- If
research_notes.mdexists, read it first - Confirm with the user:
- Total slide count (e.g., 40, 60)
- Chapter (part) count and rough proportions
- Whether to include cover / speaker intro / contact slides (default: include all)
Step 2: Write storyline.md
Construct the lecture's high-level flow narratively. Not just a table of contents — encode the logic of "why this order" so the audience journey is clear.
# Storyline: [Topic]
## Lecture Overview
- Total slide count:
- Presentation time:
- Core message:
## Overall Narrative Flow
[3-5 paragraphs describing where the audience starts, what journey they take, and what they leave with]
## Chapter Structure
| # | Type | Title | Slide count | Key role |
|---|------|------|------------|---------|
| - | Cover | | 1 | Topic and speaker intro |
| - | Speaker intro | | 1 | Build credibility |
| 1 | PART | | N | |
| 2 | PART | | N | |
| - | Contact | | 1 | Closing and follow-up |
## Key Takeaways
What the audience should remember after the lecture:
1.
2.
3.
Step 3: Write slide_deck.md
Write per-slide content. This file becomes the direct input for HTML generation.
# Slide Deck: [Topic]
---
## [Slide number]. [Slide type] — [Title]
**Section**: [section name shown in header]
**Type**: cover / speaker intro / divider / standard / contact
**Script**: [Speaker narration text — 2-4 sentences, natural spoken style. Required for every slide. Write in `script_language` from lecture-profile.md (defaults to `language`).]
### Content
- Bullet 1
- Bullet 2
- Bullet 3
### Right Panel
- Type: image / text box
- Image: [search keyword or filename]
- Text title: [if any]
- Text body: [if any]
---
Repeat this structure for every slide. Dividers include part number and description.
Korean example (actual slide content for a Korean lecture is written in Korean):
## 5. divider — The Arrival of the AX Era **Section**: PART 01 **Type**: divider ### Content - Paradigm shift in the securities industry - Definition of AX (AI Transformation)
⚠️ Uniform Layout Principle: All content slides within a single deck MUST use
standardtype. Special types (punchline,divider,profile) are reserved for their designated structural purpose only.punchlineis allowed ONLY as the last content slide with a single-statement message (≤1 line, no bullets). Mixing slide types for content slides is prohibited — each type has a distinct CSS layout (2-column vs center-aligned), and mixing breaks visual consistency.
Step 4: Balance Check
Self-review after writing both files:
- Are slide counts balanced across chapters (within ±20%)?
- Any slides with too many bullets? (4 max recommended, 5 is hard limit)
- No more than 3 consecutive slides without visuals?
- Does every slide have a
scriptfield with narration text? - If multi-language narration is requested, do translated script fields exist?
- Uniform Layout check: Do ALL content slides (between cover and closing) use
type: standard? If any content slide usespunchlineor another non-standard type, convert it tostandard.
Output Format
presentations/<project>/storyline.md— narrative flow with chapter structure tablepresentations/<project>/slide_deck.md— per-slide content with type, content bullets, and right-panel spec
Related Skills
research— producesresearch_notes.mdconsumed by this skilldesign— consumesslide_deck.mdoutput from this skill
04 同仓库更多 Skills
i18n-layout
Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.
i18n-layout
Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.
i18n-layout
Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.
i18n-layout
Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.
i18n-layout
Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.
i18n-layout
Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.