Jinchenma IP Skills
Two skills for article illustrations with a recurring character: built-in Jinchenma or your own, made from a photo
- 71 ★ on the repository
- Last commit
Extended access
Why this level
- The skill sends the article text and the character photo to the image generation service of your client
- It writes to ~/.agents and ~/.jinchenma-assets and reads pages from links you provide
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/jinchenma-ip-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skill sends the article text and the character photo to the image generation service of your client
- It writes to ~/.agents and ~/.jinchenma-assets and reads pages from links you provide
This is third-party code. Review the repository files before installing.
What it does
The repository holds two skills. The jinchenma-ip-article-illustrations skill reads a whole article, decides how many pictures are needed and which idea each serves, and draws illustrations with the same character. By default it uses a soft 3D style with a matte finish, a white background and a limited palette, and the older 2D surrealist style is switched on by an explicit request. The skill does not edit the source article, does not overwrite existing images and at the end names where to insert each one. The jinchenma-ip-builder skill creates your own character from one clear photo or a finished three-view sheet: it builds front, side and back views, writes a character-spec.md file and stores the pack in ~/.agents/jinchenma-ip-builder without overwriting existing versions. The character can then be imported into the illustration skill. There is a screenshot mode: the character explains an interface and keeps buttons and data uncovered. The screenshot stays untouched. Images are saved to .jinchenma-assets/<article>/ of the project or to ~/.jinchenma-assets/. The skill needs a client with image generation, and the README and documentation are in Chinese.
Who it is for. Article authors and bloggers who need illustrations with a recognizable character, and anyone who wants to make such a character from their own photo.
Good fit when
- A long article needs illustrations in one style with one hero, with the number of pictures set by the meaning of the text
- A character with front, side and back views is created from your own photo for continuing use
- A product screenshot is explained by the character while the interface stays untouched
- The author wants a list of insertion points: after each picture the skill names the paragraph
- The older 2D surrealist style is needed for a series started earlier and is switched on by an explicit request
Not a fit when
- You need diagrams, charts and graphs with exact data: the skill draws figurative illustrations with a character
- The agent client cannot generate images: without that the skill only prepares the concept
- You need a ready corporate brand style: there are two styles, 3D and the older 2D, and both are tied to the Jinchenma character
- You need documentation in Russian or English: the README and skill descriptions are in Chinese
Example request
Use jinchenma-ip-article-illustrations: read this article in full, draw illustrations with the character and say after which paragraph to insert each oneLimitations
- Creating pictures directly needs image generation in the client, and an article by link needs network access, otherwise paste the full text.
- Documentation and skill texts are in Chinese, and the README links an English version.
- The Builder photo is processed by your client and its generation service, whose data rules apply. The photo is not written into the character pack.
- Code, workflows and documents are MIT, while the Jinchenma images and character are CC BY 4.0 with attribution to 金尘马 or Jinchenma.
- Character packs live in ~/.agents/jinchenma-ip-builder and finished pictures in .jinchenma-assets/.
- The older 2D style is kept in the repository and is off by default.
How to disable. The skills are removed by deleting the jinchenma-ip-builder and jinchenma-ip-article-illustrations folders from the agent skills directory. Character packs sit in ~/.agents/jinchenma-ip-builder and finished pictures in .jinchenma-assets/ of the project or ~/.jinchenma-assets/, and they are deleted by hand.
FAQ
How do I install Jinchenma IP Skills?
The author advises giving an agent with terminal and file access a request to install the two skills from the repository: it picks the directory for your client itself. The README has no other commands.
Can I make a character from my own photo?
Yes, the jinchenma-ip-builder skill takes one clear photo or a finished three-view sheet and creates a character pack. The original photo is not stored in the pack, but it goes through your client service.
What is the default illustration style?
A soft 3D with a matte finish, a white background and a limited palette. The older 2D surrealist style is switched on by naming it in the request and is not remembered for the next task.
Where are pictures saved?
In .jinchenma-assets/<article name>/ inside the project, or in ~/.jinchenma-assets/ when there is no project. The skill does not overwrite finished files.
What license do the images have?
Code and documentation are MIT, while the images and the Jinchenma character are CC BY 4.0: credit 金尘马 or Jinchenma when using them.
Discussion
No questions or reviews yet.
Related
Skillsxiaohu-ip-studio: article illustrations with a fixed character
xiaohu-ip-studio
A skill that illustrates Chinese articles: it decides which sections need a picture and draws them with a recurring character
Content skills: article illustrations, covers, infographics, comics, translation, URL to Markdown and social posting
Skill that finds illustration spots in an article and generates images, diagrams and covers with three engines
Codex skills for the Xiaohuang brand character: light hand-drawn slides, article covers and web slides in one style
Eleven skills for Claude Code and Codex turn an article or video into Xiaohongshu cards, long images and a narrated short video.
A skill that turns phrases, photos, meeting notes and long texts into a monthly comic journal that grows entry by entry
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Jinchenma IP Skills (this card) | no | 71 | |
| Punk IP Illustrations | no | 260 | illustrations in a different style with a different character |
| Ian Xiaohei Illustrations | no | 13,183 | a separate illustration project with its own hero |
README in short
The README tells in Chinese about two parts of the project: creating a personal character and illustrating articles. It shows 3D examples and the preserved 2D, and describes two usage paths, a table of skills, features, installation by a request to an agent, a quick start, where results are saved, requirements, privacy and licenses.
README badge
Original SKILL.md
--- name: jinchenma-ip-article-illustrations description: 读取完整文章、Markdown、网页、帖子、单个观点或产品截图；独立导入和切换个人 IP，默认生成柔和 3D 插图，明确选择时使用旧版 2D 超现实画风，也支持截图 IP 讲解。适用于文章配图、正文插图、截图 IP 化、长文配图、人物 IP 导入与切换、article illustrations 等任务。 --- # Jinchenma IP 文章插图 ## 开始前 先完整读取共同文件，再按当前图片模式读取对应规范： - `references/ip-pack-format.md`：初始化文章人物配置，导入、激活、切换并解析当前 IP。 - `references/article-workflow.md`：确定图片数量、认知锚点、生成、保存和交付方式。 - `references/illustration-style.md`：解析本次画风与内置人物版本，只加载选中的独立风格包。默认 `jinchenma-3d` 和 3D 人物；旧版需要明确选择。 - 输入包含产品截图，或用户要求把个人 IP 融入截图时，读取 `references/screenshot-ip-mode.md`，保护截图并应用截图 IP 讲解规则。 ## 初始化并显示当前 IP 在分析文章前按照 `references/ip-pack-format.md` 完成人物解析： 1. 用户要求导入或切换人物时，先完成相应流程。 2. 第一次运行且没有文章人物配置时，在可交互场景先提示：`当前默认使用 Jinchenma。是否要更换为你自己的 IP 形象？` 用户选择更换时进入人物导入流程；选择不更换时使用内置金尘马。 3. 自定义人物只从文章 Skill 自己的 `ips/` 读取。Builder 包和外部文件只在明确导入时读取并复制。 4. 解析且只加载一个人物 IP 的 `manifest.json`，再加载本次人物版本的三视图和角色规范；同时按画风选择文件解析一个风格包。 5. 开始生图前提示：`当前使用 IP：<displayName> (<ip-id>)；人物版本：<实际版本>；插图画风：<style-id>`。单版本自定义 IP 说明使用现有参考，不虚构 3D 参考资产。 人物身份、人物参考版本、插图画风分别解析。未指定时使用本次默认 3D；截图模式保留原截图，只让当前 IP 延续选定人物版本和画风。 ## 执行工作流 1. 读取完整文章、用户提供的全部文字和明确提供的截图。用户提供文章地址时获取完整正文，不根据标题单独推测内容。 2. 提炼核心观点、关键转折、文章情绪和真正值得视觉化的认知锚点，明确每张图要帮助读者理解什么。 3. 根据正文长度和信息密度决定图片数量，合并重复内容，不按章节平均覆盖；基础状态截图能用一句话讲清楚时不选为文章插图。 4. 为每张图选择一种模式：正文内容使用文章插图模式；产品截图、运行结果和界面操作使用截图 IP 讲解模式。不得把两种模式含混地叠在一张图里。 5. 文章插图确定一个核心意思、可见场景/关系和必要物件，按本次风格包完成构图；截图 IP 讲解图锁定必须完整保留的界面区域、指向重点和安全人物位置。 6. 安排当前 IP 的叙事职责、人物尺度、动作和适合当前内容的情绪。 7. 列出少量必须准确呈现的短文字，文章插图的关键文字在正文显示尺寸下应清楚可读；截图中的原有文字、数字、按钮和界面元素全部属于受保护内容。 8. 在内部完成图片方案，不先输出方案，不等待二次确认。 9. 把当前人物 IP 的三视图作为同一角色的多角度参考，逐张独立生成或编辑，不用拼图代替。 10. 逐张检查文章含义、文字准确性与可读性、人物一致性和画面密度；截图模式额外检查界面完整性、指向准确性和关键内容遮挡。每次只修正一个主要问题。 11. 按输出规则保存图片。最终回复保持简短，按顺序列出每张图片、用途、绝对路径和建议插入位置。 ## 角色与风格边界 - 共享流程只从当前人物 IP 读取人物身份、服装、配件和稳定色。