Skip to content
Get packs

Video Generation

Lets an agent turn a text description and optional reference image into an MP4 video via Gemini Veo or MiniMax.

Source Download skill
Repository
bytedance/deer-flow
Stars
84,876
Licence
MIT
Files
3 files

What it does

Once installed, when you ask the agent for a video, it writes a structured JSON description of the scene (subject, camera, dialogue, audio), optionally uses a reference image as the first or guiding frame, and runs a bundled Python script that sends the request to a video AI service. The agent then shares the resulting MP4 file once generation succeeds. Skills are instruction packs (a SKILL.md file plus optional scripts) that teach an AI agent how to perform a specific task.

Great for

  • Creators who want short AI-generated video clips from a written scene description
  • Developers building an agent workflow that needs video output from prompts
  • Marketers prototyping short visual concepts from a product or scene brief

Skip it if

  • Users who need guaranteed, repeatable or production-grade footage: results depend on the external provider and the SKILL itself says iterative refinement is normal
  • Anyone without a Gemini or MiniMax API key: the script fails with a credentials error and produces no video
  • Generating videos of real people or recreating copyrighted scenes without rights: the example uses a film scene, and the skill gives no guardrails on this

Try asking

  • “Generate a short 16:9 video of a rainy neon city street at night”
  • “Create a video clip from this reference image showing the product slowly rotating”
  • “Make a short dramatized scene from a book's opening chapter with dialogue and ambient audio”
Step-by-step guide: how to use Video GenerationInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 3 original files, the MIT licence and an attribution note.

    video-generation.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip video-generation.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip video-generation.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Write prompts in English, as the skill instructs, and be specific about camera movement, lighting and audio.
  • Use a reference image to anchor the look and first frame, since the skill notes it improves quality significantly.
  • Check that the command succeeded before trusting output; a leftover file from an earlier run is not a new result.

What's inside

  • AGENTS.md
  • scripts/generate.py
  • SKILL.md

Redistributed unmodified from bytedance/deer-flow under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.