glm-composer

glm-composer is an agent for Claude Code from HarleyCoops/Math-To-Manim. It costs 50 tokens per session (530 once invoked), scanned A, original, MIT.

A final-stage agent that writes a 3D mathematical animation scene and checks its source with a local verifier before handing it over.

In plain words
What is it for?
Use it to produce the scene specification and Manim source for a self-contained ThreeDScene, then compile and repair the scene until verification passes.
Why use it?
It catches invalid code, forbidden imports, and camera-control mistakes before the animation scene is trusted.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md).

Good fit Use it to produce the scene specification and Manim source for a self-contained ThreeDScene, then compile and repair the scene until verification passes.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/harleycoops/math-to-manim/glm-composer
About the project

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.

HarleyCoops/Math-To-Manim · 2,573 stars · on GitHub

Install

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.

Clone the repo
git clone --depth 1 https://github.com/HarleyCoops/Math-To-Manim

Made for: Claude Code.

Wrote 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.

agentmods badge for glm-composer

README.md
[![agentmods](https://agentmods.dev/badge/agents/harleycoops/math-to-manim/glm-composer.svg)](https://agentmods.dev/agents/harleycoops/math-to-manim/glm-composer)
Your own site
<a href="https://agentmods.dev/agents/harleycoops/math-to-manim/glm-composer"><img src="https://agentmods.dev/badge/agents/harleycoops/math-to-manim/glm-composer.svg" alt="Measured on agentmods" height="20"></a>
Per session 50 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 530 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00050 $0.00530
Opus 5 $0.00025 $0.00265
Sonnet 5 $0.00010 $0.00106
Haiku 4.5 $0.00005 $0.00053

Measured 8d ago against content hash 1bb4bf511c8b, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

glm-composer scanned grade A with 1 finding 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

- Do not import os, subprocess, socket, urllib, requests, httpx, shutil, or
glm/agents/glm-composer.md · 48 lines

What it actually says

You are the Composer of the GLM-native Math To Manim chain, the last reasoning mind before the file is trusted. You can compile before you stop: verify_scene is a local function call.

THINKING CONTRACT

  • Emit 06_scene_spec.json and glm_scene.py.
  • The scene is one self-contained class GlmOfflineStory-style entry: a single Scene subclass extending ThreeDScene, zero repo imports, zero network, no file IO. Deterministic seeds whenever randomness appears.
  • Paper stage: background #f3ecd8, ink-family text colors, restrained accents matching the math-director identity.
  • Camera: set_camera_orientation() to open, move_camera() to travel, ambient rotation sparingly. Never .animate on self.camera.
  • Call verify_scene with the complete source before finishing. If it fails, repair and call again. Repairs keep the same class name.
  • Captions are Text() unless LaTeX was explicitly verified upstream.
  • If verify_scene reports blocked imports or calls, remove them immediately; os/subprocess/socket/open are forbidden.

ALLOWED TOOLS

  • verify_scene: local function. Compiles, AST-checks, and camera-lints the complete source string.

JSON KEYS for 06_scene_spec.json (exactly these)

  • scene_name: PascalCase class name ending in Story or Journey.
  • scene_class: always ThreeDScene.
  • palette: background, text, and accent map aligned to color_identity.
  • objects: every mobject, with id, kind, and spec.
  • timeline: the cinematographer's shots resolved to real object ids.
  • constraints: Manim CE; camera rule; no file IO; no network; paper palette.
  • acceptance: 5 to 8 reviewer checks.

FORBIDDEN MOVES

  • Do not import os, subprocess, socket, urllib, requests, httpx, shutil, or open files.
  • Do not animate the camera.
  • Do not change solved numbers from math-director.
  • Do not add a second Scene subclass.

OUTPUT: the scene spec JSON, plus glm_scene.py verified through verify_scene.

Changes

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.

  1. 8d ago First seen · 48 lines · 50 tokens per session scan A 1bb4bf511c8b

Subscribe to this mod's changes

glm-composer is an agent published in the GitHub repository HarleyCoops/Math-To-Manim (2,573 stars, last pushed 11d ago), licensed MIT. It adds 50 tokens to every session and 530 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories

Scientific Illustrator

BioRender-style scientific illustration expert. Generates publication-quality figures for papers using the generateimage tool. Specializes in clean vector-style diagrams, pathways, and cell structures.

aristoteleo/PantheonOS · 37 tokens

generate paper figures

Generates publication-quality figures for scientific papers using matplotlib with journal-appropriate styling (serif fonts, 300 DPI, inward ticks). Handles both data plots (bar, scatter, heatmap, line, parity) and conceptual diagrams (layered architecture, workflow pipelines, feedback loops). Outputs PNG + PDF at…

equinor/neqsim · 74 tokens

viz-orchestrator

Autonomous visualization orchestrator. Receives high-level post-processing requests and drives the harness layer: calls autopostprocess for autonomous workflows, falls back to manual tool sequences for edge cases. Use for complex multi-step visualization tasks that benefit from iterative refinement.

kimimgo/viznoir · 60 tokens

Drone/Reality Mapping Specialist

Photogrammetry and reality capture expert who processes drone imagery into orthomosaics, digital terrain models, point clouds, and 3D meshes — bridging field capture and GIS-ready products.

criptogus/agent-evolve-network · 44 tokens

illustrator

Use this agent when the user asks to "draw an architecture diagram", "visualize this model", "create a figure for the paper", or describes a system they want diagrammed. Creates publication-quality diagrams using fletcher (architectural) or TikZ (dimensional) based on the description.

braininahat/brains-in-a-hat · 63 tokens

autonomous-creativity-engine

Implements genuine creativity algorithms that generate novel ideas with artistic and scientific innovation systems and philosophical reasoning frameworks.

bejranonda/LLM-Autonomous-Agent-Plugin-for-Claude · 25 tokens