AugmentClaude
VideoWriting

Video Narration Scripter

By worldwonderer· worldwonderer/video-recap-skills· 0

Generate timestamped Chinese narration scripts for videos with validation.

Installation

  1. Make sure Claude is on your device and in your terminal.

    Skills load from ~/.claude/skills/ when Claude Code starts up — so you need it on your machine first. If you don't have it yet, install it once with the command below, then run claude in any terminal to verify.

    One-time setup
    npm i -g @anthropic-ai/claude-code

    Already have it? Skip ahead.

  2. Paste into Claude Code or into your terminal.
    Install
    Sign up to copy
    Sign up free to reveal

    This copies the whole skill folder into ~/.claude/skills/video-script-worldwonderer/ — the SKILL.md plus any scripts, reference docs, or templates the skill ships with. Safe default: works for every skill.

    Faster alternative (instruction-only skills)

    Skips the clone and grabs only the SKILL.md file. Don't use this if the skill ships Python scripts, reference markdowns, or asset templates — they won't be downloaded and the skill will fail when it tries to load them.

    Quick install (SKILL.md only)
    Sign up to copy
    Sign up free to reveal
  3. Restart Claude Code.

    Quit and reopen Claude Code (or any other agent that loads from ~/.claude/skills/). New skills are picked up on startup.

  4. Just ask Claude.

    Skills auto-activate when your request matches the skill's description — no slash command needed. Trigger phrases live in the skill's own frontmatter; you can read them in the “What this skill does” section above.

Prefer to read the source first? Open on GitHub.

When Claude uses it

Use after the video-understanding stage has analyzed a video and the narration needs writing — authors a timestamped recap script (Chinese by default), then lints and validates it into narration.json.

What this skill does

This skill authors and validates timestamped narration scripts for video recaps. It reads the video analysis (scenes, dialogue, quiet windows) and writes a structured JSON file with narration segments, timing, and metadata. The skill then lints the script against the analysis index and optionally time-aligns it to silent gaps in the original audio.

What it does

  • Authors narration.json with timestamped segments, each containing narration text, start/end times, pause duration, and whether it overlaps original dialogue
  • Optionally authors original_subtitles.json to preserve and calibrate key original dialogue spoken in gaps (with ASR corrections and proper names fixed)
  • Validates timing and structure via validate.py, checking that segments conform to the brief's scene layout and character budget
  • Aligns narration to quiet windows in full mode, rewriting segment times to match silence in the original audio
  • Supports two workflows: full-video mode (timestamps in original-video time) and orchestrated cut mode (timestamps in output timeline time after editing)
  • Reviews narration quality via an optional review.py gate that checks for hallucinations and grounds claims in visual or ASR evidence

How to use it

Read the brief at work_dir/agent_narration_brief.md to understand the video's scenes, durations, quiet windows, and character budget. Consult the raw analysis files (vlm_analysis.json, asr_result.json, timeline_fusion.json) for details. Then write narration.json as an array of segment objects with start, end, narration text, pause_after_ms, and overlaps_speech flag. Run python3 scripts/validate.py --work-dir <work_dir> --mode full to lint and align it; fix any errors and re-run until clean.

Note: this skill's internal instructions are written in Chinese. Claude reads them natively and will work with you in English.

Related skills

S

Skill Builder & Optimizer

anthropics

Create, edit, and optimize Claude skills with performance testing and benchmarking.

Official
O

Org Change Management

alirezarezvani

Guide teams through organizational changes using the ADKAR model and communication strategies.

MIT
A

Audio/Video Transcription

daymade

Transcribe audio and video files to text with fast local or remote processing.

C

Claude Export Conversation Fixer

daymade

Repair broken line wrapping in Claude Code exported conversation files.