Back to skills
SKILL.md
Web World Models
ASecurityBridge deterministic web code and generative LLMs via hybrid world models. State and physics defined by TypeScript code, LLMs generate narrative/aesthetics on top. Uses typed interfaces, deterministic hashing, graceful degradation—enabling scalable interactive environments from travel atlases to fictional worlds without databases.
- 6 stars
- 0 votes
- 0 copies
- 1 view
- Added September 9, 2026
Security analysis
100/100npx -y skills add ADu2021/skillXiv --skill web-world-models --agent claude-codeAre you the author of Web World Models?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/adu2021-web-world-models)---
name: web-world-models
title: "Web World Models"
version: 0.0.2
engine: skillxiv-v0.0.2-claude-opus-4.6
license: MIT
url: https://arxiv.org/abs/2512.23676
keywords: [world-models, simulation, language-models, web-technologies]
description: "Bridge deterministic web code and generative LLMs via hybrid world models. State and physics defined by TypeScript code, LLMs generate narrative/aesthetics on top. Uses typed interfaces, deterministic hashing, graceful degradation—enabling scalable interactive environments from travel atlases to fictional worlds without databases."
---
## Overview
Web World Models combine structured code for invariant state with LLM-driven narrative generation.
## Core Technique
**Two-Layer Architecture:**
```python
# Physics layer (deterministic code)
class Physics:
inventories: dict # Code enforces consistency
coordinates: XYZ
# Imagination layer (LLM-driven)
class Narrative:
descriptions = llm.generate(physics_state)
dialogue = llm.generate(context)
```
## References
- Physics layer as TypeScript code
- Imagination layer for aesthetics
- Deterministic hashing for consistency
Attribution
Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.