技能Skills

可以直接拿走使用的视觉方法、Agent 工具与工作流。Visual methods, agent tools, and workflows you can use directly.

Featured collection · 01

Phil Design SkillsPhil Design Skills

一套给 AI Agent 使用的视觉生成 Skills：不是收集风格关键词，而是把主题、文章、产品或现有视觉资产编译成可直接生图的完整提示词。A visual-generation toolkit for AI agents. It compiles themes, articles, products, or existing visual assets into complete image prompts instead of merely collecting style keywords.

19 套从主题出发的风格编译器，加上 1 套从 Logo、头像、角色或产品图出发的视觉身份延展工具。每套包含 Skill 指令、参考模块与可检查的输出约束；全部提示词按 gpt-image-2 编写。Nineteen style compilers that start from a theme, plus one identity-expansion tool for logos, portraits, characters, or product imagery. Each combines agent instructions, reference modules, and reviewable output constraints; all prompts are written for gpt-image-2.

Repository and installation

风格 SkillsStyle skills: 19 套真实样例19 documented styles
额外工具Additional tool: 视觉身份延展Identity expansion
目标模型Target model: gpt-image-2gpt-image-2

Generative Hover Type skillGenerative Hover Type skill

Phil Carlos

术数工程化Auditable domain rules

Numerologist SkillsNumerologist Skills

把奇门遁甲、紫微斗数与四柱八字从依赖经验的解读，转化为由固定脚本、规则集和参考资料约束的可审计工作流。Turns Qimen, Ziwei, and Four Pillars from experience-led readings into auditable workflows constrained by deterministic scripts, rulesets, and references.

Project and usage guide

其他SkillOther Skills

把领域规则、工作记录与研究方法做成可复用入口。每一项都说明它解决什么问题，并连接到可继续查看或使用的页面。Reusable entry points for domain rules, work records, and research methods. Each item states the problem it solves and links to the place where it can be explored or used.