360 Secure SkillHub
360 安全技能中心

Poster Forge

提供 AI、HTML 模板和 PIL 文本三种渲染引擎及自动回退，可按标题、尺寸预设、模板和分栏内容生成海报、社交媒体图片、封面、信息图、对比图或教程卡片。

huuuwnnn-droid
huuuwnnn-droid
热度 47.55
最近更新 2026-05-11 17:40:56
分类：创作依赖：代码执行 / 读写文件 / 第三方 API / 联网浏览搜索应用场景：图像生成 / 社交媒体 / 模板 / 演示设计

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add poster-forge
Metadata
name
poster-forge
description
Universal image and poster generator with Chinese/English text support. Use when asked to create posters, social media images, cover images, infographics, comparison charts, tutorial cards, or any visual content. Three rendering engines with auto-fallback — AI generation (Pollinations.ai), HTML/CSS templates (Chromium screenshot), and pure PIL text rendering. Supports presets for Xiaohongshu, WeChat, Instagram, Twitter, and A4 print. Triggers on "make a poster", "generate image", "create cover", "design a card", "生成海报", "做封面图", "画图".

Poster Forge

Universal poster/image generator with 3 engines and auto-fallback.

Engines

Engine How Best For Reliability
ai Pollinations.ai (flux model) Artistic/photo backgrounds ~90% (network dependent)
html Chromium headless screenshot Precise layouts, text-heavy ~99% (needs Chromium)
text PIL solid color + text overlay Simple text posters 100%

Default mode is auto: tries AI → HTML → text fallback.

Quick Start

# Simple poster with AI background
python3 scripts/generate.py --title "你的标题" --prompt "watercolor sunset" --output poster.jpg

# Split comparison (HTML engine, no network needed)
python3 scripts/generate.py --title "A vs B" --mode html --template split \
  --left-label "方案A" --right-label "方案B" \
  --left-items "优点1|优点2" --right-items "优点1|优点2" --output compare.jpg

# Tutorial card
python3 scripts/generate.py --title "教程标题" --mode html --template tutorial \
  --code-content "print('hello')" --tagline "底部提示语" --output tutorial.jpg

Parameters

Core

  • --title (required): Main title
  • --subtitle: Secondary text
  • --tagline: Bottom tagline
  • --mode: ai | html | text | auto (default: auto)
  • --output: Output file path

Templates (HTML engine)

  • --template: split | gradient | card | tutorial

Size presets

  • --preset: xiaohongshu (1080×1440) | wechat (900×500) | instagram (1080×1080) | twitter (1200×675) | a4 (2480×3508)
  • --width / --height: Custom dimensions

Split template

  • --left-label / --right-label: Side labels
  • --left-items / --right-items: Pipe-separated content items
  • --left-color / --right-color: R,G,B colors

Text overlay

  • --text-position: bottom | center | top | none
  • --font-title-size / --font-sub-size: Font sizes
  • --no-overlay: Skip PIL text overlay (when HTML already has text)

AI engine

  • --prompt: English description for Pollinations.ai

Workflow for Agents

  1. Decide content and style
  2. Choose mode: auto for flexibility, html for reliability, ai for aesthetics
  3. Run scripts/generate.py with appropriate args
  4. If using html mode with --template, add --no-overlay (text is already in the HTML)
  5. Send output image to user

Requirements

  • Python 3 + Pillow (PIL)
  • curl (for AI engine font download)
  • Chromium/Chrome (for HTML engine, optional)