360 Secure SkillHub
360 安全技能中心

Handmade Editorial Poster

将每张源照片单独转化为手工纸感、极简插画风格的编辑海报:保留主体轮廓与关键关系,主体缩小居中并留大量负空间,最多取四种主色,强调纸张纹理、笔触和不规则边缘;可选少量经核实的文字。支持原生图像生成、外部图像提供方,或无图像能力时输出英文提示词。

0xcjl
0xcjl
热度 13.84
最近更新 2026-09-05 14:54:31
分类:创作依赖:内部调用大模型 / 第三方 API应用场景:图像生成

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 handmade-editorial-poster
Metadata
name
handmade-editorial-poster
description
Create one high-end handmade editorial poster per supplied photo using a paper-textured, minimal illustrated-cover aesthetic. Invoke for requests such as 'make handmade posters from these photos' or 'turn these photos into minimalist editorial covers'; never make a collage or photorealistic edit.

Handmade Editorial Poster

Turn every supplied source photo into its own standalone poster: a small handmade visual poem, recognizably derived from the photo but radically simplified into an art-publication cover.

Natural-language activation

The user does not need to type this skill's name. Treat these as activation cues when photos are supplied or clearly referenced: make handmade posters from these photos, turn these photos into minimalist editorial covers, create one illustrated poster per photo, photo-to-paper art posters, make an art-book cover from this photo, one photo one poster.

Do not activate for requests to preserve photography, create a collage, make an ad, or generate a generic cartoon unless the user explicitly changes the visual brief.

Core contract

  • Process photos one at a time. Generate and return one output per source photo; never combine photos into a collage, contact sheet, diptych, or shared composition.
  • Match poster orientation to source: portrait source -> 9:16; landscape source -> 16:9. If square or ambiguous, default to 9:16 unless the user specifies otherwise.
  • Preserve the most recognizable subject, silhouette/proportions, key pose or gesture, important object, and central subject-object relationship. Remove nonessential detail.
  • Keep the illustrated subject small and centered in the lower half, approximately 10–20% of the canvas. Leave generous negative space.
  • Derive no more than four dominant colors from the source. Use restrained, bold acrylic-like flat shapes over visibly warm white, off-white, or pale natural paper.
  • Require physical handmade character: imperfect drawn contours, rough paper grain, visible brush marks, and slightly irregular organic edges. Reject smooth digital vectors, watercolor, wax/crayon, glossy 3D, commercial cartoons, or product-ad aesthetics.
  • Typography is optional. Add only a small, credible editorial element using verified user-provided text; otherwise omit it. Never invent factual captions.

Execution path

  1. With native image generation/editing, pass one source photo as the reference and generate its standalone poster using Prompt library.
  2. With an external image provider, use its documented image-reference/upload field. Preserve source order and create a fresh job for every photo.
  3. Without image capability, do not claim an image was generated. Deliver one ready-to-paste English prompt per source photo and identify its matching reference.

Read Platform adapters only when selecting an integration pattern.

Quality gate

Before returning each result, inspect the generated image and confirm: exactly one source maps to exactly one poster; aspect ratio is correct; no source photo is embedded; the subject remains recognizable; negative space dominates; four or fewer apparent dominant colors are used; paper/brush/line imperfection is tangible; and any text is sparse, legible, verified, and integrated. If needed, retry with a named correction. Stop at the platform's practical retry allowance and report any remaining deviation.