A sophisticated AI skill that deconstructs short videos into comprehensive narrative and psychological analysis reports for content creators and marketers.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install video-deconstruct
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install video-deconstruct using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is video-deconstruct?
video-deconstruct is a powerful tool within the Openclaw Skills ecosystem designed to transform raw MP4 videos into professional-grade analysis reports. By leveraging advanced vision and ASR models, it dissects short-form content from platforms like Douyin and Xiaohongshu into structured insights covering narrative arcs, psychological triggers, and technical execution.
This skill is particularly valuable for digital marketers and content operations teams who need to understand why specific videos go viral. It provides the ammunition needed for scripting, storyboarding, and replicating successful content patterns through a deep dive into video structure, rhythm, and audience impact.
video-deconstruct Use Cases
- Creating teardown reports for viral short-form videos on social media.
- Reverse-engineering successful video scripts and storyboards for production teams.
- Analyzing the psychological mechanisms and "hooks" behind high-engagement content.
- Extracting dialogue and subtitles automatically to assist in content localization or repurposing.
- Auditing video rhythm and BGM sync points to improve future editing workflows.
How video-deconstruct Works
- The user provides a local MP4 file path and optional flags like ASR (Speech-to-Text) through an Openclaw Skills interface.
- If the file exceeds 128MB, the system automatically uses ffmpeg to compress the video to an optimal size for API processing.
- The skill uploads the video to StepFun cloud and triggers the step-1o-turbo-vision model for visual analysis.
- If enabled, the stepaudio-2.5-asr model extracts dialogue to provide textual context to the vision model.
- The AI processes the inputs against a specialized rubric to identify themes, structures, and viral points.
- A Markdown report and a raw JSON file are generated in the output directory, and the temporary cloud files are deleted.
video-deconstruct Setup
To use this skill within the Openclaw Skills framework, ensure you have Python 3.10+ and ffmpeg installed. Set your API key and run the script:
export STEP_API_KEY=sk-your-key-here
# Basic analysis
python scripts/analyze.py /path/to/video.mp4
# Advanced analysis with Speech-to-Text
python scripts/analyze.py /path/to/video.mp4 --with-asr
video-deconstruct Data Schema & Taxonomy
The skill generates two primary files for every analyzed video, ensuring compatibility with other Openclaw Skills agents:
| File Type | Path | Description |
|---|---|---|
| Markdown Report | output/{video_stem}-report.md |
A human-readable narrative report including 10 chapters of analysis. |
| Raw JSON | output/{video_stem}-analysis.json |
Structured data for secondary processing by other AI agents or tools. |
Report Taxonomy: Includes Subject/Topic, One-sentence Summary, Narrative Description (300-600 words), Opening-Middle-End structure, Core Viral Points, and BGM Rhythm Analysis.
video-deconstruct Advanced Features
- Multi-model orchestration using StepFun's vision, ASR, and flash models for cost-effective reasoning.
- Automated ffmpeg fallback for handling large video files via intelligent compression.
- Customizable analysis rubrics via
prompts/analysis_rubric.txtto change writing styles or focus areas. - Seamless integration with StepClaw Agent framework for automated dispatching.
- Optional cloud file retention using the
--keep-uploadflag for debugging or extended processing.
SKILL.md
Loading
Related Openclaw Skills
Generate HVAC Service Company Client Education Handout
An automated tool to generate professional HVAC client education materials including visuals, FAQs, and clear service next steps.
Jike Weather Query
A comprehensive weather forecasting skill providing real-time conditions and up to 15-day extended forecasts via the Jike Data API.
Java Code Review & Business Analysis Agent
An intelligent Java auditing tool that analyzes Git changes, traces business logic via LSP, and generates comprehensive PRDs and quality reports.
Jike IPv4 Address Query
A specialized IPv4 lookup tool providing real-time geographic and ISP data for AI agents and developers.
Generate Healthcare Ad Creative Brief
An AI-driven skill designed to streamline the creation of ad creative briefs specifically for healthcare operators and clinic marketing teams.
Generate Home Remodeling Company Client Education Handout
An automated tool for remodeling professionals to generate professional, visual client education materials and FAQ guides.
Featured*
