agentleFS
wangge-aiwangge-ai/wangge-skills/skills/video-producer-core/SKILL.md

video-producer-core

Skill4 starsChanged 39 days agoMITView on GitHub

Use when planning, creating, replicating, or extending videos of any

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 8 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this as the general video production backbone.

    line 13
  2. Distinctive:

    Use this skill first for broad video requests.

    line 17
  3. Distinctive:

    If the user gives little context, proceed with sensible defaults and state assumptions.

    line 29
  4. Distinctive:

    Write a one-sentence thesis.

    line 43
  5. Distinctive:

    Keep one main visual idea per beat.

    line 49
  6. Very specific:

    Use references/production-template.md for the brief, folder structure, and storyboard columns.

    line 50
  7. Distinctive:

    Use original or user-owned assets.

    line 54
  8. Distinctive:

    Keep narration, ambience, transition hits, and effects as separate source files before final render.

    line 59
  9. Distinctive:

    Put source assets, generated assets, audio, captions, and renders in separate folders.

    line 63

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Video Producer Core

Overview

Use this as the general video production backbone. It handles the production brief, script, storyboard, asset plan, audio plan, build plan, and verification, while specialized skills handle domain-specific visuals such as Manim technical diagrams or platform-specific styles.

Route the Request

Use this skill first for broad video requests. Then add specialized skills only when needed:

NeedAdd
Technical diagrams, formulas, data flows, system architecturemanim-hyperframes-video
HyperFrames HTML composition, subtitles, transitions, renderhyperframes:hyperframes and hyperframes:hyperframes-cli
Raster images, B-roll plates, cover imagesimagegen
Website capture into videohyperframes:website-to-hyperframes
Remotion/React video compositionremotion:remotion-best-practices

Minimum Inputs

If the user gives little context, proceed with sensible defaults and state assumptions. The minimum useful brief is:

  • Topic or product
  • Target duration
  • Audience
  • Style reference or mood
  • Voiceover choice: no voice, temporary TTS, user-provided audio, or final TTS
  • Output: script only, storyboard, project files, or final MP4

For stronger results, also ask for required talking points, forbidden claims, brand rules, aspect ratio, target platform, and reference links/files.

Workflow

  1. Frame the video
    • Write a one-sentence thesis.
    • Pick the format: explainer, tutorial, product demo, course, ad, social clip, recap, or documentary.
    • Choose aspect ratio and duration before writing the timeline.
  2. Write script and storyboard
    • Split into beats of 5-20 seconds.
    • Keep one main visual idea per beat.
    • Use references/production-template.md for the brief, folder structure, and storyboard columns.
  3. Plan visuals
    • Classify each beat as footage, screen recording, AI image/B-roll, motion graphic, chart, diagram, or title card.
    • Use original or user-owned assets. Do not depend on copyrighted source clips unless the user explicitly owns or licenses them.
    • Record prompts for generated assets in the project.
  4. Plan audio
    • Generate temporary TTS when the user has not provided narration.
    • Keep narration, ambience, transition hits, and effects as separate source files before final render.
    • On Windows, check SAPI voices as a fallback if model-based TTS fails.
  5. Build
    • Put source assets, generated assets, audio, captions, and renders in separate folders.
    • Build a short draft first for timing and style.
    • Expand to the full target duration after the draft checks out.
  6. Verify
    • Confirm the final file has video and audio streams with ffprobe.
    • For HyperFrames, run npx hyperframes lint, validate, and inspect.
    • Extract representative frames and inspect title, dense middle scenes, subtitle scenes, and ending.

Environment Preflight

For production work, run the bundled script when available:

powershell -ExecutionPolicy Bypass -File "scripts/check_video_environment.ps1"

It checks common dependencies such as FFmpeg, Node, Chrome, Python, Manim, MiKTeX, and Windows TTS voices.

Deliverables

A complete video job should leave behind:

  • script.md or script_full.md
  • storyboard.csv
  • assets/ai/ with prompts recorded
  • assets/audio/ with source audio
  • composition source files
  • renders/*.mp4
  • verification notes: duration, resolution, fps, audio stream, and warnings

Common Mistakes

  • Starting composition before the script and storyboard are clear.
  • Treating a style reference as permission to copy assets.
  • Putting the whole video into one tool when different beats need different tools.
  • Forgetting audio verification and delivering a silent MP4.
  • Trusting render success without checking representative frames.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.