菜单
Generate YouTube chapter markers from transcript content.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
/blog-schedule - Manage Post Schedule
Generate a blog post from source materials (text, PDFs, images, transcripts). Use when user wants to create a blog post or article.
Render carousel slides from social/carousel.md into 1080x1350 PNG images.
Push blog posts to Google Docs for review, then pull comments and suggestions back as feedback for discussion.
Generate AI images for content using Gemini or other image generation APIs. Use when user needs diagrams, illustrations, or visual content.
Initialize a new content project with the standard directory structure in projects). Use when user wants to create a new blog post and one doesn't exist for it.
| name | video-chapters |
| description | Generate YouTube chapter markers from transcript content. |
Generate YouTube chapter markers by reading the transcript and identifying topic boundaries. No script — do this by reading the transcript and reasoning about where topics change.
All outputs go in <project_root>/claude-edits/.
<project_root>/claude-edits/
<video_stem>_chapters.txt # YouTube-format chapter list
<video_stem>_chapters.json # Structured chapter data with timeline times
_analysis.json or _review.md)Chapters reference source timestamps, but the edited video has cuts. To map:
# Build map from EDL kept segments
timeline_pos = 0.0
for seg in kept_segments:
if source_time >= seg["start"] and source_time <= seg["end"]:
timeline_time = timeline_pos + (source_time - seg["start"])
timeline_pos += seg["end"] - seg["start"]
_chapters.txt (paste directly into YouTube description):
0:00 Introduction
0:42 The Just File
1:48 Running Scripts with AI Agents
...
_chapters.json (for Resolve markers):
{
"chapters": [
{"source_time": 0.0, "title": "Introduction", "timeline_time": 0.0},
{"source_time": 51.3, "title": "The Just File", "timeline_time": 42.0}
]
}