historical-science-video-prod

Automated Generation Skill for Historical Science Popular Science Short Videos

Task Objective

• This Skill is used for: automatically generating the full-process materials for a 3-minute popular science short video on historical science topics
• Capabilities include: spoken narration script creation, storyboard script planning, Veo2 prompt generation, character image design
• Trigger condition: the user needs to produce a historical science popular science video and hopes to quickly obtain a complete material package

Prerequisites

• Dependency description: no additional installation is required; dependency packages are already declared in the dependency field
• Input requirements: the user provides the popular science topic, era background, and core scientific conclusion

Operation Steps

• Standard process:
  1. Prepare input information
     • Determine the popular science topic (such as "The Origin of the Butterfly Effect", "The Discovery of Penicillin")
     • Clarify the era background (such as "1961", "1928")
     • Extract the core scientific conclusion (such as "sensitivity to initial conditions")

1. Execute the generation script

   • Call scripts/generate_video_materials.py to process...
   • Pass in parameters: --theme, --era, --core_conclusion, --output_dir

2. View output materials

   • output/script.txt - 3-minute spoken narration script (about 900-1100 words)
   • output/storyboard.md - Storyboard script table (30-35 storyboards)
   • output/veo2_prompts.txt - Veo2 prompt list
   • output/character_design.md - Character image specifications

• Optional branches:
  • When the style needs to be adjusted: modify the retro style parameters in references/style-guide.md
  • When the script structure needs to be optimized: adjust the time allocation in references/script-structure.md

Resource Index

• Required script: see scripts/generate_video_materials.py(Purpose and parameters: generate the full-process material package, receiving theme, era, and core conclusion parameters)
• Domain reference: see references/multi-agent-architecture.md(When to read: to understand the multi-agent collaboration system architecture)
• Domain reference: see references/style-guide.md(When to read: when adjusting retro style parameters)
• Domain reference: see references/veo2-prompt-template.md(When to read: to understand the standard format of Veo2 prompts)
• Domain reference: see references/script-structure.md(When to read: to understand script structure and time allocation)

Notes

• The script will automatically call the large model to generate content; generation takes about 1-2 minutes
• The output files will be saved to the specified output_dir directory
• Ensure that the provided core scientific conclusion is accurate, as the script will generate the narration based on it
• The generated storyboard script strictly follows a 3-minute duration, with each shot lasting 4-6 seconds

Usage Example

• Function description: generate a video material package for the topic "The Origin of the Butterfly Effect"
• Execution method: the script automatically calls the large model for generation
• Key parameters or guidance points: theme, era, core conclusion
• Simple example code or command:

python scripts/generate_video_materials.py \
  --theme "蝴蝶效应的起源" \
  --era "1961年" \
  --core_conclusion "初始条件的微小变化会导致巨大差异，无法长期预测天气" \
  --output_dir "./output"