1 Modellix

modellix-infographic Skill

为单张信息图、视觉总结或信息图示规划一条提示词。当用户想要信息图、可视化总结,或需要把素材浓缩成一张对比图、流程图、漏斗图、循环图、仪表盘或模块总览图时,使用这个技能——哪怕他们只说"把这些内容变成一张图"。确认结构、宽高比和画面语言,引用素材中的事实,然后交给 modellix-design 生成。跳过社交平台卡片、普通海报,以及只需执行现成提示词的请求。

安装方式:把技能目录放入 ~/.claude/skills/(Claude Code)或在 claude.ai 设置中启用;也可复制右侧安装命令一键添加。

查看源码

技能指令原文(SKILL.md)

Modellix Infographic

Turn a source into one prompt for one infographic. This skill plans the image.
modellix-design generates it.

Do not submit, poll, download, authenticate, or inspect model schemas here. Do not
run the Modellix CLI. Do not recommend a model, a provider, or a slug. When the
user wants the image made, finish the brief first and then hand the prompt to
modellix-design.

A cover, thumbnail, or carousel for a named network belongs to
modellix-social-card. A poster, product shot, or illustration that is not an
information diagram belongs to modellix-image-prompt-engineering. Files already
on disk belong to modellix-ffmpeg or modellix-moviepy.

Route the request

Use this skill when the user wants one image that explains a source: a
comparison, a sequence, a funnel, a cycle, a stack of layers, a hub, a metric
board, or a set of peer modules.

Ask once for the source when the user names the format but not the content. Do
not invent facts to fill an empty request.

Skip or keep the pass short when:

  • The user already supplied a finished prompt and asked only to run it. Pass that

prompt to modellix-design.

  • The request is a social card for a specific network. Use modellix-social-card.
  • The request is an article, blog, or newsletter cover. Use modellix-cover-image.
  • The request is a set of illustrations placed through an article. Use modellix-article-illustrator.
  • The request is a knowledge comic, educational comic, biography comic, or four-panel strip. Use modellix-comic.
  • The request is a poster or product image with no structured facts. Use

modellix-image-prompt-engineering.

  • The request is a commerce UGC video that still needs a shot plan. Use

modellix-ecomm-ugc-video.

  • The request is video editing, credentials, task status, or a download.

Plan the image

Read references/structures.md to choose the structure. Read
references/prompt-shape.md before writing the prompt.

  1. Copy the facts that must appear. Keep every number, name, and quotation. Drop

credentials, tokens, and secrets. Do not add a statistic the source does not
contain.

  1. State one purpose: what the viewer should understand in a glance.
  2. Split the source into sections. Each section has a short headline, the exact

text, and a visual role such as a bar, a step, or a spoke.

  1. Offer two or three structures that fit, each with a one-line reason. A

structure the user already named is the confirmation. Otherwise wait. Also
confirm the aspect and the on-image language. Language follows the source
unless the user names another.

  1. Write one prompt for one flat image. Quote the exact words, say which line is

largest, and describe where each section sits. Tell the model to copy the
quoted text and not add labels of its own.

Do not ask for SVG, HTML, or a second pass that paints text onto the bitmap.
Crowded copy is fixed by shortening the quoted labels, then regenerating.

Brief

Return this brief:

Infographic brief
- Purpose:
- Structure:
- Aspect:
- On-image language:
- Sections:
  1. Role / exact text / visual role
- Prompt:
- Next step: stop after prompt | pass the prompt to modellix-design
- Model: <only a slug the user already named | omitted>

Leave Model blank unless the user already named a slug. Do not browse the
catalog from this skill.

If the user asked only for a prompt, stop after the brief. If the original
request asked for the image, pass the prompt to modellix-design and let that
skill ask for a model, then own schema validation, credentials, paid submission,
waiting, and download.

Revision

When a generated infographic misses, change the prompt and return a new brief:

  • Wrong or extra number: quote the source figure again and say not to invent one.
  • Garbled label: shorten the quoted text and repeat the copy instruction.
  • Wrong reading order: restate the confirmed structure and where each section sits.
  • Too much text: move nuance out of the image and keep the labels.

modellix-design decides whether an existing task can be recovered or a new
submission is appropriate. This skill does not cover the bitmap with corrected type.