Video Explainer logo

Video Explainer

Community
runesleo
video-explainer

Turn a research brief, finished script, or generated MP4 into a reviewed and verified vertical explainer video. Use when an agent should choose the simplest current production strategy, preserve strict final-artifact quality gates, run the deterministic Remotion fallback, diagnose the local toolchain, or verify an existing MP4 before any manual publication step.

Overview

Publisherrunesleo
Repositoryclaude-video-kit
Skill namevideo-explainer
Stars
122
Forks
26
Bundled files
7
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 7 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by runesleo on GitHub. Read the source before you install it.

Installation

Install the Video Explainer AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/runesleo/claude-video-kit/tree/main/skills/video-explainer
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/runesleo/claude-video-kit.git /tmp/claude-video-kit
mkdir -p .claude/skills
cp -r /tmp/claude-video-kit/skills/video-explainer .claude/skills/video-explainer
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Video Explainer in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Video Explainer on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Video Explainer is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Video Explainer

Version note: v0.3.0-rc.1 / 2026-09-27 / Make production strategy adaptive while keeping the deterministic Remotion recipe as a strict fallback. Final-artifact acceptance is strategy-independent and bound to the exact MP4 bytes. Keep upload, account access, credentials, private voice assets, hosted rendering, and publication out of scope.

Produce a local video and evidence packet. Never upload or publish from this skill.

Core contract

This Skill separates what must be true of the final artifact from how the artifact is produced.

Before forcing an existing recipe, run a capability challenge:

  1. Can the current model or available capability produce the final MP4 directly at the required quality?
  2. Which requirements are real product, safety, brand, or evidence invariants?
  3. Which existing steps only compensate for older model limitations?
  4. Can intermediate stages be skipped without weakening final acceptance?
  5. What is the simplest strategy likely to pass the final gates?
  6. If that strategy fails, fall back to the deterministic repository-owned recipe.

Choose one execution strategy:

  • direct_artifact — a current model/capability produces the final MP4 directly. Do not require motion-storyboard.json, script.json, TTS, alignment, or Remotion merely to satisfy a historical recipe.
  • agent_orchestrated — an agent may choose and combine tools as implementation details. Intermediate artifacts are optional unless they are needed to satisfy the brief or evidence requirements.
  • legacy_remotion — use the repository-owned deterministic path when direct generation cannot preserve facts, branding, layout control, narration identity, or other required qualities.

All three strategies converge on objective MP4 verification plus an independent final-artifact review. Production freedom never bypasses final acceptance.

Install

The deterministic fallback orchestrates a local clone of claude-video-kit; it does not bundle the renderer. From the repository root, install the runtime and register the Skill with:

bash
npm ci --prefix remotion
npx skills add runesleo/claude-video-kit --skill video-explainer

For repository development, run commands from the repository root.

Workflow

  1. Read the brief and run the capability challenge above. Select direct_artifact, agent_orchestrated, or legacy_remotion. Prefer the simplest strategy that can plausibly pass the same final acceptance contract.

  2. If using direct_artifact or agent_orchestrated, produce the final 9:16 MP4 without manufacturing legacy intermediate files just to satisfy the old recipe. Then skip to Step 6.

  3. If using legacy_remotion, run the read-only doctor before changing a project:

    bash
    node scripts/video-explainer.mjs doctor --output <output-directory>

    Read setup-and-doctor.md when doctor blocks or the environment is new.

  4. For legacy_remotion, turn the brief into motion-storyboard.json before treating the script as ready. For every scene, define its information purpose, visual metaphor, motion grammar, why motion communicates better than a static card, duration hint, and explicit 3D decision.

    Read motion-storyboard.md, then validate:

    bash
    python3 scripts/check_storyboard.py <project>

    Production planning in this deterministic path must not default to title + card + bullets. Generic entrance/exit effects (fade, slide, zoom, etc.) do not count as an information model by themselves. 3D is an exception that must earn its render/debug cost with a concrete comprehension benefit.

  5. For legacy_remotion, convert the approved information plan into script.json, require the independent pre-render review, and render only from a current pass receipt:

    bash
    node scripts/video-explainer.mjs review <project> --input <project>/review-input.json
    node scripts/video-explainer.mjs render <project>

    Read brief-to-script.md and review-gate.md. The pre-render gate checks facts, structure, duration, visual feasibility, hierarchy, simplicity, clarity, legibility, craft, privacy, and copyright. Any script edit makes that receipt stale.

    On macOS, use --demo-quality only for a credential-free first success. It forces the built-in demo voice and script-timed captions; that explicit demo path may omit a storyboard and is never publication-quality. Read render-and-verify.md before running a custom project.

  6. For every strategy, verify the exact final MP4 objectively:

    bash
    node scripts/verify-shorts.mjs <video.mp4>

    For the deterministic Remotion path, prefer bound render metadata when available:

    bash
    node scripts/verify-shorts.mjs <project>/out/full.mp4 \
      --metadata <project>/metadata.json

    The objective verifier checks canvas, duration, and scene rhythm. It does not prove factual correctness, visual polish, privacy, copyright, narration consistency, or end-card correctness.

  7. Require an independent final-artifact review bound to the exact MP4 SHA-256:

    bash
    node scripts/video-explainer.mjs review-render <video.mp4> \
      --input <render-review-input.json>

    The final-artifact review is strategy-independent. It requires:

    • facts
    • structure
    • duration
    • hierarchy
    • simplicity
    • clarity
    • legibility
    • craft
    • privacy
    • copyright
    • audio_consistency
    • end_card

    Record production_strategy in the review input when known. It is traceability metadata, not a quality waiver.

    If any final-artifact check is fix or block, revise or regenerate. A changed MP4 invalidates the receipt.

  8. Before any distribution handoff, finish all cover/subtitle/audio edits and follow the repository's final release gate: python3 scripts/video_release_gate.py evaluate, then verify on the final master. Require the platform's independent video_final_review.v1, three hash-bound QA docs, current quality profile, and objective audio/cover checks. Keep the Step 7 final-artifact review; it cannot substitute for the platform release receipt. build-distribute-pack verifies that receipt before writing any handoff. Downstream publishers must verify again immediately before an authorized upload/schedule action.

  9. Report the chosen production strategy, video path, objective verification result, final-artifact review receipt, release-gate status (or explicitly not run), elapsed time, and known limitations. For legacy_remotion, also report storyboard validation and the pre-render receipt. A local pass is evidence for the artifact, not publication approval.

Fallback rule

If direct_artifact or agent_orchestrated fails final acceptance, first determine whether another autonomous attempt can reasonably fix the defect. If deterministic control is required for facts, data visualization, brand layout, narration identity, or reproducibility, fall back to legacy_remotion.

Do not force the deterministic recipe merely because it exists. Do not accept a direct artifact merely because it was generated by a newer model.

Canonical first success

Run the repository-owned neutral demo:

bash
node scripts/video-explainer.mjs demo --output /tmp/video-explainer-first-success

The demo exercises the deterministic fallback. It must create a current pre-render pass receipt before TTS, render locally without an API key on supported macOS, and run the shorts verifier. The explicit demo-quality script-alignment path may omit motion-storyboard.json; this exception exists to preserve a credential-free tooling smoke test, not to define production quality. The demo intentionally does not self-approve the final-artifact gate because an independent reviewer must inspect the actual MP4. It must not read an account, upload media, or send the brief to a remote service.

Stop conditions

  • Stop the legacy_remotion production path when motion-storyboard.json is missing or fails its gate. The only exception is the explicit demo-quality script-alignment path.
  • Stop legacy_remotion before TTS/render when the pre-render review is missing, stale, fix, or block.
  • Stop every strategy short of publication-quality acceptance when final-artifact review is missing, stale, fix, or block.
  • Stop when objective MP4 verification fails.
  • Stop distribution handoff when final release evidence is missing, stale, malformed, or fails evaluate / verify.
  • Stop when doctor reports a required runtime missing for the deterministic fallback; give its exact action.
  • Stop before any upload, public post, deploy, paid API call, credential setup, or account action and return that boundary to the user.
  • Read errors-and-privacy.md for recovery and data-handling rules.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Video Explainer AI skill do?

Turn a research brief, finished script, or generated MP4 into a reviewed and verified vertical explainer video. Use when an agent should choose the simplest current production strategy, preserve strict final-artifact quality gates, run the deterministic Remotion fallback, diagnose the local toolchain, or verify an existing MP4 before any manual publication step.

Why use Video Explainer on TypingMind?

Because you install it once and use it with any model. Video Explainer is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Video Explainer in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/runesleo/claude-video-kit/tree/main/skills/video-explainer. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Video Explainer?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Video Explainer?

As many as you like. As long as a model supports skills, you can use Video Explainer with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Video Explainer AI skill free?

Yes. It is published on GitHub by runesleo under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇