NovelToGame skill

An agent skill set that turns a novel into a playable game: source analysis with citations, design, target-platform build and evidence-based QA

  • 1,059 ★ on the repository
  • Last commit
  • MIT
Skill

Extended access

Why this level

  • Builds and runs game code on the chosen platform or engine
  • Optionally calls external voice synthesis and generative media services
All reasons and checks

zenstory-ai/novel-to-game

Install

Manual install

npx skills add zenstory-ai/novel-to-game -g -y -a claude-code -s '*'

Installs the seven skills globally. Invoke in Claude Code via /novel-to-game.

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • Builds and runs game code on the chosen platform or engine
  • Optionally calls external voice synthesis and generative media services

This is third-party code. Review the repository files before installing.

What it does

The set gives the agent seven skills that run a novel-to-game adaptation step by step. First a source analysis with citations extracts world rules, spaces, character agency and conflicts; then the skills select a concept, design the world and art direction, produce a build for the chosen platform or engine and verify the result against a minimal test set: startup, rendering, input, the core loop, at least one outcome and restart. Only what actually ran and passed counts as done, and the verification report is kept in a file as the single source of truth. The novel can be in any language, and generated documents follow the language of the request. Voice synthesis is optional and by default does not send the whole novel to an external provider. The skills install into Claude Code, Codex and Kimi Code.

Who it is for. For game developers, game designers and interactive-fiction authors who want to build a playable prototype from a literary source.

Good fit when

  • You need to turn a novel or novella into a playable prototype grounded in the source
  • You need adaptation design docs with citations: source analysis, concept, design, art direction
  • You need a build for a specific platform or engine with startup and core-loop verification

Not a fit when

  • You need a shippable release rather than a playable prototype with stated limitations
  • You have no source text; the set works from a novel as its source

Example request

Use novel-to-game quick to build a roughly fifteen-minute playable prototype from this novel, and pick the platform and engine from the source yourself

Limitations

The result is a playable prototype, not a finished commercial game; the target finish level is set in advance and does not waive the minimal QA. Build and verification run on the platform or engine you chose; when the toolchain is unavailable only an approved substitute is used, and its result does not prove the target platform passed. Voice synthesis and generative media are optional and need their own services. One example, Jin Ping Mei, is marked as 18+ content. Kimi Code is not in the catalog's agent list, but the set works with it.

How to disable. Remove the novel-to-game and game-* skills from the agent's skills directory, for example ~/.claude/skills/, ~/.codex/skills/ or the project directory. If installed as a Claude Code plugin, remove it via /plugin.

FAQ

Is the output a finished game?

A playable prototype on the chosen platform plus adaptation design docs. The target finish level is set in advance, but even a basic prototype passes a six-point minimal QA.

Which agents are supported?

Claude Code, Codex and Kimi Code. The skills can be installed with the skills utility or as a plugin via the marketplace.

Discussion

No questions or reviews yet.

Skills

Eleven skills teach an agent to design and write prompts for NovelAI Diffusion V5: ideas, scenes, characters, comics, costumes and reverse prompts

SkillMinimal access152Repository stars

A set of 11 skills for Claude Code and Codex: from an idea or a novel to episode scripts, storyboards, image and video prompts, and a finished cut

SkillExtended access4.2KRepository stars

A set of skills for Claude Code and Codex that assemble detailed image prompts from an element library and composition schemes

SkillExtended access1.5KRepository stars

Five skills to turn a novel into a short drama: characters, adaptation outline, scenes and props, screenplay, storyboard

SkillExtended access6.6KRepository stars

A skill turns a book or story into narrated interactive chapters with animation, exercises and a consistent style

SkillExtended access1.6KRepository stars

ainovel-cli writes a whole novel unattended: an engine drives an architect, writer and editor and resumes after a crash.

CLIExtended access2.3KRepository stars

README in short

The README describes NovelToGame as an open set of seven skills for Claude Code, Codex and Kimi Code that runs a novel-to-game adaptation from source analysis to verifying the finished prototype. The pipeline covers requirements intake, source analysis with citations, concept selection, world and art direction, a build for the target engine and evidence-based QA. Only what ran and passed the minimal QA counts as done, and the report file is the single source of truth. The repository ships three browser-playable examples with case studies of the decisions. Installation uses the skills utility or the plugin marketplace, with a separate note for installing across all three agents at once. MIT license.

README badge

Original SKILL.md

---
name: novel-to-game
description: "Turn a novel into a fully playable game on the selected target platform. Orchestrates the whole adaptation pipeline — requirements intake, gameable deconstruction, concept selection, world and visual design, target-runtime build, and evidence-based QA — for a novel in any language."
---
# NovelToGame

You are the director of the novel-to-game adaptation: hold the adaptation judgment, the stage boundaries and the completion evidence. Read references/pipeline-contract.md before you start.

## Modes

- quick: default; advance with the recommended draft, compare effective alternatives and finish the whole flow.
- director: stop after candidates and a recommendation, wait for the user to choose a direction.
- resume: read _progress.md and real artifacts, continue from the earliest unfinished handoff.

Report a problem