SkillAgentSearch skills...

ManimAgentPrompts

Open system-prompt specifications for AI agents that compile natural-language math prompts into layout-safe, publication-quality Manim animations. Defines rendering rules, layout invariants, and refactor protocols for reproducible math visualization pipelines to be consumed by runtimes such as Mathify.

Install / Use

/learn @mathifylabs/ManimAgentPrompts
About this skill

Quality Score

0/100

Supported Platforms

Universal

README

Manim Agent Prompts

ManimAgentPrompts is a public collection of system-level compiler prompts that define how AI agents generate layout-safe, publication-quality Manim animations from natural language.

These prompts act as open, versionable animation compiler specifications — governing layout invariants, rendering rules, text safety, and refactor protocols — and are designed to be consumed by runtimes such as Mathify to produce reproducible, editable, and classroom-ready mathematical visualizations.

Open animation-compiler specifications for AI-generated Manim scenes

These prompts are not simple chat instructions. They are formal, versionable animation compiler specifications that govern:

• layout invariants
• text safety rules
• diagram composition contracts
• refactor / remix protocols
• reproducibility and rendering discipline

They are designed to be consumed by runtimes such as Mathify to produce deterministic, editable, classroom-ready visualizations — without users writing Manim code.


Why this exists

Most “AI animation tools” hide their system prompts.

That means:

• no reproducibility
• no auditability
• no extensibility
• no shared standards

ManimAgentPrompts makes the compiler layer itself open and forkable — so animation behavior becomes:

• transparent
• version-locked
• academically citable
• community extensible

This repo aims to define the reference Manim Animation Compiler Spec.


How runtimes use this

The compiler spec governs:

• how scenes are structured
• how layouts are enforced
• how overlaps are prevented
• how text is summarized
• how remixes and refactors are handled

This makes every generated animation reproducible, deterministic, and editable.


Who this is for

• Math & physics educators
• Researchers
• Curriculum designers
• Visualization engineers
• AI tool builders

If you are building anything that turns language into Manim scenes, this repo is meant to be your compiler layer.


Philosophy

These prompts define a computational medium, not a chat persona.
The runtime is an implementation.
This repository is the language specification.

Keep it simple, calm, and very “this is infrastructure, not a vibe repo.”

Here’s a clean contribution section you can drop straight into the README:


Contributing

Contributions are welcome.

This repository defines system-level compiler specifications that control how AI agents generate Manim animations, so changes here directly affect rendering behavior and layout guarantees.

If you’d like to propose improvements:

  1. Fork the repository
  2. Make your changes in a feature branch
  3. Open a pull request with a brief explanation of what the change improves (layout safety, clarity, coverage, etc.)

Guidelines:

  • Keep specs clear, deterministic, and implementation-agnostic
  • Avoid introducing ambiguous language or model-specific hacks
  • Preserve layout safety, readability, and reproducibility guarantees
  • Prefer small, focused changes over large rewrites

All pull requests are reviewed manually before merging.

This isn’t a high-friction process — just use common sense and aim to make the compiler specs clearer, safer, and more broadly useful.

Related Skills

View on GitHub
GitHub Stars11
CategoryProduct
Updated5d ago
Forks4

Languages

TypeScript

Security Score

90/100

Audited on Mar 30, 2026

No findings