Adaptation skill pack
Novel to Game
Adapt a novel into a playable game build grounded in the source, and check it running on the platform you choose.
813 GitHub stars7 skills
npx skills add zenstory-ai/novel-to-game -y -gWhat it is
Novel to Game is a seven-skill workflow for Claude Code, Codex and Kimi Code that adapts a novel into a playable game candidate. It builds a cited source bible, compares adaptation concepts, designs gameplay and art direction, and builds for the target runtime you choose. A QA contract records evidence that the build launches, renders, takes input, completes its core loop, reaches an outcome and restarts.
What you need
- LicenseMIT, open source.
- HostAny one of these agents:Claude CodeCodex CLIGoogle AntigravityOpenCodeZCodeOpenClawReasonixKimi CodeDeepSeek Harness
Start in 3 steps
- Install
npx skills add zenstory-ai/novel-to-game -y -g - Entry commandRun
/novel-to-gamein your agent host. - Follow a guideFind your first task under “How it works” below and open its guide.
Writing craft
56 craft articles, each built on a method file in this project's skills. Start with the complete workflow, or open a task.
Complete workflow
Build a playable game prototype
By creative task
How it works
- Can I make an interactive story rather than a combat game? Yes: set a narrative-led experience in the brief and concept. Decisions can change testimony, what characters know and later scenes; they need not be disguised as cards, turns or an affinity meter.
- How do choices change more than the wording? Make important actions readable before selection and show their consequences in scene prose, not only state labels. The original ferry example writes advance information, a separate consent exchange and three closing passages for two decision points. Expressive wording can remain flavor; do not promise another ending or a hidden score for every line.
- What may I change from the novel? Keep cited source facts and uncertainties in
analysis/SOURCE_BIBLE.md. Separate approved adaptation additions from evidence; do not let implementation silently rewrite the agreed brief or source bible. - How much should I adapt first? Choose a source range explicitly, then one player role, a concrete objective and an outcome that preserves the intended experience. A small build does not excuse missing source coverage for a requested whole-book adaptation. The worked slice-selection brief compares a summary, a bounded ferry decision and an expanded town scope; game-world-design then specifies the selected slice before full art production.
- What do I receive, and what remains unproven? The quick start explains the brief, design files and built candidate. Design-only work is not a playable game. The six runtime checks cover the recorded run; they do not establish every branch, subjective quality or commercial readiness.
- How should I revise my first game candidate? Describe the exact choice or scene, your input, what you expected and what appeared. The ferry candidate-feedback examples distinguish a misleading refusal label, a displayed ending versus a wrong state transition, and an author-requested rule change. Keep approved facts and untouched design decisions; a scoped revision is not a reason to regenerate the whole concept.
What makes it different
- Source evidence and adaptation decisions are written down as explicit files rather than an unsupported genre guess.
- The target platform, engine and test runtime are recorded; passing in a substitute runtime does not establish target-platform compatibility.
- The repository links to example games built with this workflow; you can play one before you start.
Who it is for
- Authors and IP holders who want an interactive edition of a story
- Game prototypers
- Narrative designers
Terms it uses
source bible whitebox QA contract
Sources
Based on the source as of 2026-09-29; links point at that version.
- Seven skills and installation
- Language, product scope and target-runtime contract
- Build-tool limits and truthful completion reporting
- Six runtime checks and limits of QA
- Causal agency, consequence resolution and narrative-led design
- Knowledge boundaries, choice effects and honest convergence
- Required narrative design handoff and smallest playable slice
- Player verbs and rule-resolved state commits