agentleFS
LaplacianAILaplacianAI/manimvox/skills/manimvox-creative/SKILL.md

manimvox-creative

Skill2 starsChanged 21 days agoMITView on GitHub

The brief, theme, story and narration for manimvox videos: asking the person what they want (audience, takeaway, feeling, their angle, how it opens and ends) into BRIEF.md, choosing a direction (three storylines, one look, a transition per chapter seam, a BANNED list) into DIRECTION.md, choosing a theme from the request and topic and mapping it onto the kit's colour roles, story shapes for a paper, a concept, a playlist or a code change, narration style, defining each word once, the facts discipline (FACTS.md), and writing and critiquing the script (opening, bridges, ending, pace) before any scene. Use before any outline, when setting the series.json theme, outlining a video or series, writing the script, or writing narration.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    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:

    Do these in order, and settle each before the next, because a change late is expensive: a new font changes every text width, and a new …

    line 11
  2. Very specific:

    Facts: Read facts before writing a line of narration.

    line 17
  3. Distinctive:

    Narration: Read narration when writing the voice.

    line 19
  4. Distinctive:

    If you cannot ask the user a question, state the choice you made and continue.

    line 24
  5. Distinctive:

    Never invent a number, name, path or endpoint.

    line 26

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

manimvox creative

The look, the shape and the words of a video. Each is decided from the request and the subject, never from a preset: manimvox ships no themes and no house style.

Do these in order, and settle each before the next, because a change late is expensive: a new font changes every text width, and a new outline changes every scene.

  1. Brief. Read brief first. Before any outline, ask the person who the video is for, what they should take away and feel, their own angle, and how it should open and end, in two short rounds of questions with suggested options. Write the answers to BRIEF.md; decide what they leave open and mark it (chosen by the agent); manimvox check brief must pass. With no one to ask, decide on purpose and list your choices in the report. Everything below follows from the brief. 0c. Direction. Read direction after the brief. Write three storylines, each a device (what the story is) paired with a form (how the pictures and seams carry it), with the exact first 2 seconds and the last shot; recommend one; save them to DIRECTION.md. Then add the look from one medium, motion by role, the opening shot, the ending, a transition for each chapter seam, varied title entrances, and a BANNED list. The person picks; with no way to ask, take the recommendation and name the other two.
  2. Theme. Read theming first. It gives the order to look in (the user's request, the subject's brand, the subject's own domain colours, then mood from topic and audience, then the kit's neutral dark default), the table that maps what you find onto the kit's colour roles, how to handle fonts and a logo, and manimvox check theme, which must pass before any scene is written. It allows light grounds and textures as well as dark ones. It ends with two worked examples, one in the spirit of 3Blue1Brown and one for a product with a brand file.
  3. Story. Read story when outlining, after DIRECTION.md. Each kind of video has a shape: a paper (hook, claim, evidence, limits, takeaway), a concept (a question, intuition built visually, the name, the formula last, where it shows up), a playlist (ground, mechanisms, assembly, edges) and a code change (why, what changed, before and after, risk, how to try it). It also covers the running example, one idea per chapter, and the beats table that pairs each idea with the picture that carries it, and each video with the question it opens on and the question it hands to the next.
  4. Facts. Read facts before writing a line of narration. Every claim comes from FACTS.md, one claim per line with its source, labelled [SOURCE], [BUILT], [DESIGN], [OPEN] or [EXTERNAL]. A design is never called shipped; an invented example is marked on screen.
  5. Script. Read script before writing any scene. Write every spoken line in episodes/epNN_name.md: an opening and an ending that follow DIRECTION.md (the usual shape is a problem and one question, answered at the end), and chapters that end on a takeaway and a bridge. Critique it against its checklist, in writing, until every answer is yes. Only then write code.
  6. Narration. Read narration when writing the voice. Spoken English, at most 45 words a line, numbers as they are said, no hype, each term defined once in a phrase, lines of 15 to 30 words, and word-count targets for each running time, planned at 2.6 words a second.

Rules that hold everywhere

  • The user's explicit request beats every default here. Where the request is silent, follow the references and say in your report which choice you made and why.
  • If you cannot ask the user a question, state the choice you made and continue; do not stall.
  • Downloading a font, a model or a package is a download: ask first. With no way to ask, the answer is no, and you use fonts already on the machine.
  • Never invent a number, name, path or endpoint. If a scene needs an illustrative example, mark it as one on screen.
  • Pronunciation is fixed in the voice, not the text: see manimvox-voice.

Related skills

  • Writing the scenes, the kit's components, the loop of draft, look, render: manimvox-core.
  • Command output, including what manimvox check theme prints: manimvox-cli.
  • Voices, engines and consent for online or cloned voices: manimvox-voice.

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.