SkillFoxx

nai5-prompting skill

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

  • 152 ★ on the repository
  • Last commit
  • GPL-3.0
Skill

Minimal access

Why this level

  • The skills only describe a workflow and output text, and run nothing and call nothing
  • The composed prompts are meant for NovelAI, where the user follows the service rules and age restrictions
All reasons and checks

miint-sunny/nai5-promptingAuthor:Miint-Sunny

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/nai5-prompting

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Minimal accessThe tool runs nothing and makes no network calls.

  • The skills only describe a workflow and output text, and run nothing and call nothing
  • The composed prompts are meant for NovelAI, where the user follows the service rules and age restrictions

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

What it does

nai5-prompting is a set of Agent Skills for composing NovelAI Diffusion V5 prompts. The main form is 11 separate skills with `SKILL.md` files and references: ideas and story seeds (`nai5-ideas`), scene and setting (`nai5-scene-design`), style, light and color (`nai5-art-direction`), original character design (`oc-character-design`), illustration composition, comic storyboards, costume design, shared writing syntax, turning a comic or a costume into a prompt, and a reverse prompt from a finished picture (`nai5-reverse-prompt`). The agent takes only the skill it needs and reads its references as the task requires. If you ask for a single main string, it returns a `Base Prompt:` block without character lines, and when editing words it changes only the fragments you named and lists side effects separately. The skills produce prompt text and run nothing: you get an image by pasting the prompt into NovelAI by hand. Besides the split version the repository holds archives: a merged one (two volumes), a single file, a short `nai5-mini` skill for ordinary illustrations and variants for Gemini's web interface, where several skills cannot be installed.

Who it is for. Anime illustrators and comic authors working in NovelAI who want an agent to come up with the idea and write the V5 prompt itself.

Good fit when

  • You only have a theme and a couple of images and the agent must work out the scene and composition
  • An original character needs looks, clothing and behavior with steady details across a series of frames
  • A multi-panel comic storyboard turns into main and character lines of a prompt
  • You need a prompt from a finished picture that you can edit, or just a part of it
  • One word is corrected in a finished prompt without touching the rest of the text

Not a fit when

  • You need automatic image generation: the skills write text, and calling NovelAI takes separate tools
  • A model of another generation or another service: the rules are written for Diffusion V5
  • You want a catalog of other people's prompts: this is a method, not a collection
  • You want a short memo for one picture: the author keeps a separate `nai5-mini` skill for that

Example request

Come up with a rainy street scene with two heroines and write a NovelAI V5 prompt, with the main string and the character lines kept separate

Limitations

All material is in Chinese, references included, so a non-Chinese reader needs an agent that translates tasks and answers. The GPL-3.0 license sets its own terms for copying and distribution. Gemini on the web needs the strict version, and the strict-safe variant, which the author tested only for import and not for image generation, covers safety blocks. The skills do not create images and do not fill in the NovelAI interface, so that is done by hand.

How to disable. Delete the nai5-* and oc-* folders from the agent's skills directory (`~/.claude/skills/`, `~/.agents/skills/` or `.cursor/skills/`) and restart the client.

FAQ

How do I install nai5-prompting in Claude Code?

Copy the 11 folders from the repository's `skills/` into `~/.claude/skills/` or a project's `.claude/skills/` and restart the client. Easier still, send the agent the repository link and ask it to follow INSTALL.md, and it lays out the files itself.

Which version of the set should I pick?

The split version of 11 skills: it is the main one and installs by default. Take the merged or single-file version only if your client accepts one skill, and the short `nai5-mini` for ordinary single illustrations.

What do I do in Gemini's web interface?

Import the strict package: all 11 skills are merged into one with `SKILL.md` at the archive root. If Gemini blocks it on safety grounds, take strict-safe, named `nai5-art-studio-safe`.

Does the set create images?

No, it produces only prompt text and ideas. You get the picture by pasting the prompt into NovelAI. Autofilling the interface and calling the service take other tools.

Does the set work for Cursor and Codex?

Yes. INSTALL.md gives `~/.cursor/skills/` for Cursor and `~/.agents/skills/` for Codex, plus project paths. The structure of a skill folder with SKILL.md inside has to stay intact.

Discussion

No questions or reviews yet.

Skills

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

SkillExtended access1.5K

A skill that searches a library of 10,000+ image prompts, shows sample images and tailors the chosen prompt to your article or script

SkillMinimal access1.4K

Nano Banana Pro Prompts Recommend

AI Image Prompt Recommender

A skill that picks ready image-generation prompts from a large library, shows sample images and assembles a prompt for your text

SkillExtended access1.9K

Seven Claude Code skills and a prompt library for the Seedance, Kling and HappyHorse video models: prompts, storyboards, debugging

SkillMinimal access450

A Claude Code skill that turns an idea into a five-stage prompt for Sora, Kling, Veo and Seedance, with templates for 21 genres

SkillMinimal access531

Five image-generation skills: Qwen-Image prompts, logos with shared geometry, diary comics, transparent PNGs and process dioramas

SkillMinimal access241

README in short

The README explains that the project gives an agent a method for writing NovelAI V5 prompts and runs nothing. Three ways to install the skills are described: hand the agent the link, download an archive and give it to the agent, or copy the folders by hand. Tables of versions follow (split, merged, single, short, strict) together with the list of 11 skills with their references and archives.

README badge

Original SKILL.md

---
name: nai5-writing
description: Use when the user needs general NovelAI prompt syntax, field or preset guidance, wording for a settled ordinary illustration, or an exact edit or correction to existing prompt text.
---

Copyright (C) 2026 Miint-Sunny · GPL-3.0-only。项目与公开来源：[nai5-prompting](https://github.com/Miint-Sunny/nai5-prompting)。完整许可见 [LICENSE](LICENSE)，贡献范围见 [NOTICE](NOTICE)。

# 通用提示词写法与精确改词

承接用户已定内容，按 [写法契约](references/conventions.md#writing-conventions) 选择本次终点；共同语法、预设、用户事实与原串保护只在此确立。普通完整画面再读 [普通写法](references/illustration.md)，用户限定了修改范围（一处或多处）时按 [精确修改](references/conventions.md#writing-exact-edit) 处理：范围外字符不动，连带改动只列不改，范围不清先问。技术问题按需读 [共享依据](references/sources.md) 的对应节；历史依据仅在用户明确要求核对时读取。

Report a problem