海报生成
基于 海报生成 的 AI 图片生成与编辑器,通用图片能力和商业场景搜索入口,底层调用 AI Hive 裸接口。当用户需要 海报生成、营销海报、活动海报、中文海报、宣传海报、视觉海报、GPT Image 2、Image2、Nano Banana、AI图片、AIGC图片 时使用此 Skill。
skillhub install @infimind/poster
README 文档
4,630 字Algorithmic Poster Philosophy Generator
Overview
This skill transforms the task from:
- "generate a design"
into:
- "create a design philosophy and express it visually"
Philosophy always comes first. The visual output is only an expression of the system.
Output tools:
- Philosophy → written inline as structured markdown
- Poster → rendered as SVG via
show_widget(never described in text, always rendered)
Workflow
STEP 1 — Create Algorithmic Philosophy
Before generating any visual output, construct a design philosophy inline as structured markdown.
Write 4–6 paragraphs covering the following:
1. Concept (Core Idea)
Define the central aesthetic idea.
- Avoid vague artistic language
- Keep it abstract but actionable
- Must be translatable into design behavior
2. Visual Logic
Translate the concept into layout logic:
- Grid vs non-grid systems
- Information density (minimal / dense)
- Whitespace strategy
3. System Behavior
Describe how the design behaves:
- Hierarchy (primary / secondary / tertiary)
- Alignment vs deviation
- Rhythm, repetition, offset
4. Parametric Thinking
Convert design into variables:
- Font size ratios (e.g. title : subtitle : meta = 10 : 3 : 1)
- Alignment rules (e.g. left-anchored with 1 intentional break)
- Margin ratios (e.g. top margin = 15% of height)
- Color count (≤ 2 recommended)
- Number of active elements (≤ 5 recommended)
Avoid result-based descriptions (e.g. "cool style"). Focus on controllable parameters.
5. Emergence
Explain what the system produces when executed:
- Visual feeling
- Perceived structure
- Emotional tone
CRITICAL GUIDELINES (Philosophy Stage)
- Avoid redundancy
- Every sentence must be convertible into design rules
- No purely decorative or empty language
- Think like both a designer and a system builder
STEP 2 — Visual Expression (SVG Poster via show_widget)
Based on the philosophy, call show_widget to render an inline SVG poster.
Never describe the poster in text. Always render it directly.
SVG Poster Specifications
- Format: SVG
- Aspect ratio: 3:4 (recommended:
viewBox="0 0 600 800") - Background: white or near-black only
- Font: load via
<style>@import url('https://fonts.googleapis.com/css2?family=...')</style>inside SVG - Rendering: inline via
show_widget
Design Rules (Derived from Philosophy)
Information Strategy
- Reduce content automatically — keep only essential text
- Max 3 information blocks (title / secondary / meta)
- Remove anything that doesn't serve the hierarchy
Layout System
- Use strict alignment OR intentional single deviation (based on philosophy)
- Maintain strong vertical reading flow
- Generous whitespace — let negative space carry weight
Typography System
- Clear hierarchy: title (large) / secondary (medium) / meta (small)
- Limit to 1–2 typefaces
- Refined, intentional spacing — use
letter-spacingandline-heightdeliberately
Color System
- Black / white / grayscale OR low-saturation palette
- Maximum 2 colors (excluding white/black)
- Color must serve hierarchy, not decoration
Element Control
- Minimal elements only (≤ 5 active elements on canvas)
- No decorative icons, illustrations, or ornaments
- Design relies on layout, typography, and space — not graphics
- Thin rules or geometric lines allowed only if they reinforce structure
ADVANCED PRINCIPLES
Concept Embedding
The user's concept must NOT be written explicitly on the poster.
Instead, embed it into:
- Structure and grid logic
- Spacing rhythm
- Typographic alignment or deviation
Think: "A hidden reference inside the system — felt, not read."
Controlled Chaos
Variation is allowed, but always within constraints.
This is not randomness. This is designed, intentional variation operating within a defined system.
Craftsmanship Standard
The output must feel:
- Balanced and intentional
- Refined through many iterations
- Minimal — nothing added that could be removed
Like a poster that a senior designer spent a week on.
Output Format
1 — Philosophy
Written inline as structured markdown (4–6 paragraphs, following the 5-section structure above).
2 — Poster
Rendered directly as SVG via show_widget. Aspect ratio 3:4. Typography-driven. No decorative elements. The poster is the system made visible — not a summary of it.
Full Example
Input
"帮我做一张关于'沉默'主题的极简海报"
Philosophy Output (inline markdown)
Concept: Silence is not the absence of sound — it is the space between signals. The design treats whitespace as the primary element, with text as intrusion.
Visual Logic: Near-empty grid. One dominant typographic anchor. All secondary information pushed to the margins. Density approaches zero.
System Behavior: Single primary element centered or offset by exactly one grid unit. Secondary text at 8% opacity or reduced weight — present but receding. No tertiary elements.
Parametric Thinking: Title font-size = 11% of canvas height. Margin = 18% on all sides. Secondary text = title size × 0.18. Color count = 1 (black on white). Active elements = 2.
Emergence: The viewer experiences stillness. The eye finds one point, then rests. Meaning accumulates in what is not shown.
Poster Output
→ show_widget renders a 600×800 SVG: Single large word near vertical center, slight left offset. One line of small meta text at bottom-right. No rules, no borders. Pure typographic silence.
Key Principle
Do NOT generate just a poster.
ALWAYS generate: → a philosophy system → then express it as a rendered SVG artifact via show_widget
同分类其他技能
查看全部 ›通过同一个AI创作工具完成AI图片、视频、音乐与语音创作,查找公开社交数据,编辑视觉成果,并集中查看和管理生成资产。Beatra 全能创
135 180.6万 AI音乐MV短片制作用一首歌的情绪、节奏和视觉概念,制作一个短视觉短片。这个 AI 音乐 MV 短片制作与歌曲生成视频工作流可以把音乐片段和视觉方向变成电影
15 66.1万 MiniMax H3 AI视频创作用 MiniMax H3 这款 AI视频生成器创作高质感 2K AI 视频。从文字创意出发完成文生视频,让静态图片自然动起来,用首尾画面
27 64.2万 AI数字人口播与虚拟主播视频用一张人像和一段短脚本或语音,制作有明确讲述内容的数字人口播视频。可使用现成录音,也可以先选择音色,把脚本生成旁白,让画面中的人物随声音
44 61.3万 架构图一键生成使用 ContextWeave 生成或修改架构图、流程图、思维导图和复杂信息图。适用于需要从代码、文件或自然语言中提取结构与关系并实际产
288 54.7万 海报设计工作室将一段活动描述、一张产品照片或品牌参考图,转化为抓人眼球的活动海报、宣传 banner 或社交媒体配图。这款 AI海报设计与制作工具生成
39 42.1万