cs-narration-phrase-timeline SKILL.md skillA

cs-narration-phrase-timeline from chenshuo2004/cs-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add chenshuo2004/cs-skills/cs-narration-phrase-timeline@git:20261006.c2a8989

Pins git:20261006.c2a8989 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
chenshuo2004/cs-skills
Path
cs-narration-phrase-timeline/SKILL.md
Versions recorded
1
Latest version
git:20261006.c2a8989, committed Oct 6, 2026
Size
4,027 bytes
Repository stars
204
License
MIT

Use when the user needs a Chinese narration phrase timeline from real audio: phrase-timeline.json, 逐句或逐短语时间标注, Whisper word-boundary alignment, timed captions, or narration-driven infographic and motion-graphics timing. Requires the final narration audio and the exact spoken copy; do not use for estimated timing, generic editing, or MP4 rendering.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

<!-- CS Skills · 陈硕 | portable skill entry | https://github.com/ChenShuo2004/cs-skills -->

# 口播短语时间表

## 目标与边界

将**真实口播音频**与实际朗读的中文稿对齐,交付按短语排序的 `phrase-timeline.json`。它为字幕、重点词、动态图文和 Motion Graphics 提供真实的毫秒与帧号边界。

这个 Skill 不生成配音、不按字数估算时间、不编辑时间线,也不导出 MP4。没有最终旁白音频时,不能把计划时长伪装成对齐结果。

## 输入与交付

需要:

- 最终旁白音频;若是音色克隆工作流,使用克隆完成后的音频,不使用参考音频。
- 与音频实际朗读内容一致的完整口播稿;不能用分镜提纲、摘要或事后改写稿替代。
- 可选:现有任务目录、目标帧率(默认 30 fps)和输出位置。

交付:

- `alignment.tokens.json`:真实语音识别的词级边界。
- `phrase-timeline.json`:短语到真实毫秒/帧的映射。
- 对齐结果摘要:音频总时长、全文覆盖率、失败原因或可用文件位置。

字段与验收要求见 [时间表结构](references/schema.md)。

## 工作流

1. **先找已有产物。** 如果项目任务目录已经有 `phrase-timeline.json`,先检查它是否来自最终音频且符合结构;核对音频与稿件的哈希或等效来源记录以及目标帧率;一致且有效时直接复用,避免重新识别。仅文件名相同不足以证明仍有效。
2. **选择正确入口。** 目标项目已有“动态信息图”或同类工作台时,优先让它在配音完成后自动生成时间表。只有用户已提供最终音频、需要独立产物时才走手动对齐。
3. **验证实现,不猜命令。** 手动执行前,确认目标项目实际存在音频对齐器、短语构建器、Node、FFmpeg/FFprobe 与项目 Python 环境。常见实现会有 `align.mjs` 与 `semantic_timeline.py`;不要假定某个旧的 CLI 文件仍存在。
   选择 WhisperX / faster-whisper 后,先核对所需 ASR 与中文对齐模型的版本、缓存路径、完整文件与哈希,复用有效缓存;只下载缺失文件,不为每个 Skill 新建模型副本。让实际加载器使用同一缓存或检查结果的本地模型目录。其他对齐后端沿用其配置,不强制下载 Whisper 模型;VAD 与 Python/GPU 依赖另行检查。
4. **先对齐音频,再构建短语。** 音频对齐器先写词级时间戳;短语构建器再使用原文和这些时间戳生成结果。不要按字符数量、页面时长或预设语速补全缺失时间。
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add chenshuo2004/cs-skills/cs-narration-phrase-timeline@sha256:1b8157f7d7cfd19a

The line at the top of this page pins the label git:20261006.c2a8989, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_f6bhr55dfxfwlfka.svg)](https://markdownregistry.com/a/art_f6bhr55dfxfwlfka)

1 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 1 version of this file, committed on Oct 6, 2026.

versioncommittedcommitsizeaudit
git:20261006.c2a8989 latest2026-10-06 c2a8989 4,027 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (4027 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

chenshuo2004/cs-skills · 204 stars · license MIT · pushed 2026-10-08 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_f6bhr55dfxfwlfka
GET https://markdownregistry.com/api/v1/resolve?ref=chenshuo2004/cs-skills/cs-narration-phrase-timeline
GET https://markdownregistry.com/api/v1/blob/1b8157f7d7cfd19a4d025dedc0ec444f6b44f6eed2068079c2e482750462ebb7

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from chenshuo2004/cs-skills

AGENTS.md AGENTS.md file
chenshuo2004/cs-skills · AGENTS.md
git:20261001.0b27495 · audit A · 204 stars
cs-auto-videl SKILL.md skill
chenshuo2004/cs-skills · cs-auto-videl/SKILL.md · Use when the user types $cs-auto-videl or asks for Douyin/TikTok ecommerce video workflows, including competitor…
git:20261006.c2a8989 · audit A · 204 stars
cs-chatcut-video-blueprint SKILL.md skill
chenshuo2004/cs-skills · cs-chatcut-video-blueprint/SKILL.md · 将自媒体素材、内容想法或已有口播写成中文剧本与制作蓝图,附逐镜分镜、逐图图片提示词和按需的视频动作提示词。用于选题、剧本、分镜与制作前文本交接;不操作项目,不生成或导出媒体。
git:20261006.c2a8989 · audit A · 204 stars
cs-chatcut SKILL.md skill
chenshuo2004/cs-skills · cs-chatcut/SKILL.md · 当用户要规划一条 ChatCut 短视频,或需要从内容想法到首个 ChatCut 项目的零到一操作指引时使用:先选择最值得做的主题和唯一核心观点,再完成可确认的中文口播稿、素材清单、Motion Graphics…
git:20261006.c2a8989 · audit A · 204 stars
cs-checkpoint-version SKILL.md skill
chenshuo2004/cs-skills · cs-checkpoint-version/SKILL.md · Use when the user asks to save the current repository version, create a rollback point, checkpoint dirty worktree…
git:20261006.c2a8989 · audit A · 204 stars
cs-clean-code SKILL.md skill
chenshuo2004/cs-skills · cs-clean-code/SKILL.md · 整理代码、重构、审查实现、修复已定位问题或同步开发文档时使用。聚焦需求与业务行为、最小修改和本地验证;Git 或部署收尾由 cs-ending-time 处理。
git:20261006.c2a8989 · audit A · 204 stars
cs-code-story-video SKILL.md skill
chenshuo2004/cs-skills · cs-code-story-video/SKILL.md · 把中文口播、角色和参考素材制作成可验收的 16:9 代码叙事视频:A-roll 角色讲述、全屏场景 B-roll、真实音频时码、Remotion 样片与成片、3:4/4:3 杂志风封面。仅用于代码驱动的叙事视频,不替代 ChatCut…
git:20261006.c2a8989 · audit A · 204 stars
cs-code-video SKILL.md skill
chenshuo2004/cs-skills · cs-code-video/SKILL.md · 用代码做动画视频:页面暴露 render(t),Playwright 逐帧截图 + FFmpeg 合成 mp4,可接合成配乐、出图和配音。通用模式先明确主题、规格、风格和声音,出风格卡、分镜与关键帧后逐镜制作;Vibe知识大赏模式额外查证来…
git:20261006.c2a8989 · audit A · 204 stars
cs-codex-image SKILL.md skill
chenshuo2004/cs-skills · cs-codex-image/SKILL.md · 通过 Codex 内置生图生成或编辑图片,并将该能力接入 Claude Code、Claude 桌面版或其他本地 MCP Agent。用于 cs-codex-image、内置生图、不用 API Key 出图、Claude 调用 Codex…
git:20261006.438ac88 · audit A · 204 stars
cs-digital-human-product-video-pipeline SKILL.md skill
chenshuo2004/cs-skills · cs-digital-human-product-video-pipeline/SKILL.md · 当用户要把产品事实、口播、数字人、产品界面和 CTA 制作成可验收的数字人产品介绍视频时使用:统一预检 ChatCut、FFmpeg、ComfyUI、Fish/TTS 与 Remotion,按 plan、sample、batch…
git:20261006.c2a8989 · audit A · 204 stars
cs-ending-time SKILL.md skill
chenshuo2004/cs-skills · cs-ending-time/SKILL.md · 完成已实现功能的验证、Git 提交/推送、PR 或部署交付。按用户在会话中已授权的动作执行,核对仓库、提交与真实部署结果;单纯本地收尾不自动公开发布。
git:20261006.c2a8989 · audit A · 204 stars
cs-frontend-design SKILL.md skill
chenshuo2004/cs-skills · cs-frontend-design/SKILL.md · 设计、实现、迭代或评审真实前端页面与交互流程,包括 SaaS、仪表盘、编辑器、落地页和可视化工具。关注产品任务、现有设计语言、可达状态与浏览器验证;不用于纯后端工作。
git:20261006.c2a8989 · audit A · 204 stars

Every file in chenshuo2004/cs-skills

Browse by kind, by grade A, or by owner.