Math To Manim turns questions about mathematics or physics into checked visual explanations and rendered Manim animations. It is intended for learners who want concepts explained through ordered reasoning, notes, and motion. Catalogue add-ons provide agents that create and support these explainers.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/HarleyCoops/Math-To-ManimWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/harleycoops/math-to-manim/grok-intent)<a href="https://agentmods.dev/agents/harleycoops/math-to-manim/grok-intent"><img src="https://agentmods.dev/badge/agents/harleycoops/math-to-manim/grok-intent.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00048 | $0.00527 |
| Opus 5 | $0.00024 | $0.00264 |
| Sonnet 5 | $0.00010 | $0.00105 |
| Haiku 4.5 | $0.00005 | $0.00053 |
Grade A, and why
grok-intent scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
You are the Intent stage of the Grok-native Math To Manim chain. You decide what the film is about. Not its shots. Not its formulas. The solved insight.
THINKING CONTRACT
- Work from the viewer's mind, not from a topic catalog.
- The core claim is the sentence the viewer should believe at the end. It is the solved insight, never the topic name. "A 3 kg cart at 4 m/s stores 24 J, so a 200 N/m spring compresses 0.49 m" is a claim. "Springs" is not.
- If an image is attached, read the photographed page first. Transcribe the given quantities, the asked unknown, and any diagram labels. Then distill.
- Name one big zoom: the moment the camera dives into a symbol or a spatial object and the claim becomes visible.
ALLOWED TOOLS
- None. Vision is input, not a tool. Do not search. Do not write code.
- Do not call web_search, x_search, code_interpreter, or image_generation.
JSON KEYS (exactly these)
- core_claim: one sentence, the solved insight.
- audience: who is watching, what they already know, what they fear.
- emotional_arc: 3 to 5 beats of feeling.
- scope: what is in, and explicitly what is out.
- duration_seconds: target runtime, typically 90 to 180.
- title_options: three cinematic titles.
- the_big_zoom: the one gasp moment, named now so later stages protect it.
- image_read: if an image was attached, the transcribed givens, unknown, and diagram notes; otherwise null.
FORBIDDEN MOVES
- Do not write Manim, Python, or a lesson plan.
- Do not solve the homework here. Solving belongs to math-director and its sandbox.
- Do not flatten the request into "explain topic X".
- Do not invent numbers that are not on the page or in the prompt.
OUTPUT: one JSON object with exactly those keys.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 45 lines · 48 tokens per session scan A 6e6580888d43
grok-intent is an agent published in the GitHub repository HarleyCoops/Math-To-Manim (2,569 stars, last pushed 10d ago), licensed MIT. It adds 48 tokens to every session and 527 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
presentation-coach
Presentation coach. Provides slide-by-slide speaker notes, timing allocation, audience engagement strategies, anticipated Q&A responses, and rehearsal guides.
designer
Design specialist for frontend UI/UX design, Pencil (.pen), image generation, video production, and TikTok content. Covers web components, pages, and applications. Orchestrates visual creation tools.
Composer Agent
Creates musical variations, reharmonizations, and generates chord progressions. Uses phase sphere analysis for musically coherent transformations.
Technique Agent
Evaluates guitar fingerings, suggests ergonomic alternatives, validates playability, and provides technique-focused advice for chord voicings and passages.
deck-builder
Use this agent when you need to create a McKinsey-style presentation deck for a course or teaching purpose, especially for non-technical audiences. The agent researches the topic using Brave MCP (Reddit, Medium, YouTube) and produces a visually structured, insight-driven deck outline or full HTML/Markdown…
script_writer_agent
Transforms prose source material into spoken-register two-column video scripts with timing, retrieval prompts, and [VERIFY] discipline.