dacnay816y62-hub GitHub avatar

fantasy-movie-poster-skill

dacnay816y62-hub

A skill package that analyzes a film’s genre, audience, mood, and motifs, then generates a layered Chinese movie poster workflow with base, typography, and grid composite.

Stars

17

7-day growth

No data

Forks

0

Open issues

0

License

No data

Last updated

2026-07-27

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It goes beyond simple ‘image + title’ by requiring deep film analysis and a structured four‑layer process, enforcing genre fidelity and avoiding common poster clichés like fake laurels or random quotes.

Who it is for

  • Film marketers and distributors
  • Graphic designers specializing in movie posters
  • AI art practitioners exploring structured prompt workflows
  • Independent filmmakers needing low‑cost poster concepts

Use cases

  • Create a 9:16 Chinese poster for a new film release
  • Develop layered poster templates for iterative design
  • Generate typography layers with image models instead of local fonts
  • Validate poster assets and composite them with grid‑based positioning

Strengths

  • Explicit film‑type analysis before design (genre, audience, mood, visual motif)
  • Separation of base layer and typography layer for fine‑grained control
  • Scripts for compositing, validation, and quality checking
  • Clear rules to preserve genre promises (e.g., not darkening a Pixar‑like film)

Considerations

  • Experimental and personal package; no formal license for public reuse
  • Works only for 9:16 Chinese posters (vertical aspect ratio)
  • Requires additional image‑generation models for the typography layer; local text replacement is discouraged

README quick start

Fantasy Movie Poster Skill

中文电影海报生成与设计指导 skill,用于把影片关键词、故事方向、剧照、三联分镜或视觉参考转译成可执行的电影海报方案与分层生成流程。

核心目标不是“剧照加标题”,而是先判断电影,再设计海报:明确类型、观众、情绪、视觉母题、标题构成和底图承载面。

What It Does

  • 进行电影海报前置分析:类型、亚类型、受众、情绪温度、核心冲突、视觉母题。
  • 根据参考图提取视觉 DNA,而不是机械复刻参考图。
  • 支持 9:16 中文电影海报方案、提示词、质量检查和分层生成工作流。
  • 支持“底图层 / 字体图层 / 网格合成”的可复用流程。
  • 强制避免真实电影节桂冠、真实片商标识、随机媒体引语和无意义小字。
  • 所有虚拟制作署名统一使用“梵想美学”。

Layered Poster Workflow

默认推荐分层流程:

  1. LLM 前置分析
    先判断这是一部什么电影。尤其遇到三联参考图时,要判断它是合家欢动画、文艺悬疑、运动压力片、东方权谋、科幻隔离,还是其他类型。

  2. Base Layer
    生成或裁切无字底图。底图必须提供文字可以“附着”的物理表面,例如幕布、玻璃、墙面、天空、桌面、地面反射、门框、头盔、球场、物件表面。

  3. Typography Layer
    字体图层也由图像模型单独生成,不用本地普通文字排版冒充字体设计。中文主标题必须清晰可读。

  4. Grid Composite
    本地只做裁切、缩放、透明度、混合模式和网格定位,不重新打字。

Important Rule

如果参考图明显是儿童 / 合家欢 / Pixar-like 动物 IP 气质,不要把它硬改成阴郁空场文艺片。可以高级化,但不能背叛影片类型承诺。应保留角色魅力、表演能量、温暖色彩和家庭观众可进入性。

如果不能一模一样复刻参考图,就不要做“像但不准”的中间态。要么精确复现,要么做清晰的主题衍射。

Examples

LiteraryEpicSports
InstitutionalSuspensePoster

Repository Structure

fantasy-movie-poster/
  SKILL.md
  manifest.json
  references/
    layered-cover-workflow.md
    genre-presets.md
    typography-system.md
    composition-models.md
    quality-control.md
  scripts/
    build_prompt.py
    composite_layers.py
    validate_project.py
    validate_poster_text.py
    validate_skill.py
  templates/
  assets/
  examples/

Scripts

Composite a base layer and a typography layer:

python scripts/composite_layers.py `
  --base path/to/base.png `
  --type path/to/type.png `
  --out path/to/composite.png `
  --grid-out path/to/grid.json `
  --canvas 1024x1792 `
  --type-box 0,0,1,1 `
  --mode screen `
  --opacity 0.95

Validate the skill package:

python scripts/validate_skill.py

Design Notes

  • 默认画幅:9:16。
  • 中文标题优先,英文副标题只做辅助节奏。
  • 标题可以成为主视觉,占据 30%-60% 的画面,但必须服务叙事。
  • 底图应低到中等对比,避免脏黑、重 HDR、恐怖片化和无意义颗粒。
  • 不自动保留三联画结构,除非用户明确要求。

License

Personal / experimental skill package. Add a formal license before public reuse if needed.

Related repositories

Similar projects matched by category, topics, and programming language.

lopopolo
Featured
lopopolo GitHub avatar

harness-engineering

Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.

AI & Machine LearningAI Agents
2,390
slvDev
Featured
slvDev GitHub avatar

esp32-ai

A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.

AI & Machine LearningLarge Language Models
1,960
littledivy
Featured
littledivy GitHub avatar

mimic

mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.

AI & Machine Learning
1,482