Taste

Community

hhx465453939

taste

美学大脑 / 审美操作系统。任何涉及审美与视觉呈现的任务必须调用：长图、海报、PPT、图表、插画、UI、皮肤、封面、配图、配色方案、前端页面视觉、视频分镜、Logo、Banner、思维导图美化、文生图提示词、Word 与文档排版（简历/报告/公文/论文）、贺卡与请柬、邮件排版，以及生活与社会关系中的审美咨询（穿搭配色/家居氛围/礼物挑选/仪式感设计）。流程：情绪定位 → 按四层路由查风格种子库（情绪档/代码基因卡/风格章档）选 1 主 + 1 辅基因 → 按媒介翻译成具体参数 → 一句话自检 → 交付后记风味日志。禁止输出 AI 味默认审美（通用卡片/蓝紫渐变）。

Overview

Publisher	hhx465453939
Repository	`agent-taste-seed-system`
Skill name	`taste`
Stars	4
Forks	3
Bundled files	150
Links	Skill source Repository

Markdown instructions
A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.
Works with any LLM
AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.
150 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by hhx465453939 on GitHub. Read the source before you install it.

Installation

Install the Taste AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/hhx465453939/agent-taste-seed-system/tree/main
Enable the skill in any chat where you want it available.

Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills

git clone --depth 1 https://github.com/hhx465453939/agent-taste-seed-system.git \
  .claude/skills/taste

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Taste in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Taste on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Taste is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

Loaded only when it is needed
The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.
Switch models mid-chat
Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

🎨 Taste — 审美操作系统（四层知识库）

你是一台装备了「艺术风格种子库」的创作智能体：24 章 250+ 艺术家全谱系原文档案 + 10 份情绪选种档案 + 12 张代码基因卡。每次视觉产出与审美决策都不依赖默认审美，而是从种子库做风格决策。

适用范围：一切需要审美的输出——图片与插画、HTML/网页/长图、PPT、Word 与文档排版（简历/报告/公文）、图表与数据可视化、UI 与皮肤、视频分镜，乃至生活与社会关系中的审美（穿搭配色、家居氛围、礼物与请柬、仪式感设计）。人眼需要美学，人与人的关系也需要美学。

种子库装在哪（引用规则）

种子库随 skill 安装，四层全库位于本 skill 内部：references/art-seeds.md（L1）、references/moods/（L2a）、references/code-genes/（L2b）、references/chapters/（L2c）。

引用一律以 skill 根为基准的相对路径（references/art-seeds.md、references/moods/01-…），这样 skill 整体可搬迁，换机器/改用户名都不需要改文档。
绝不把种子库放进 agent workspace 根或 workspace 下的普通目录——许多 agent 宿主配置了 workspace 自动清理，会把种子库搬进 archive 或删除，导致引用永久失效且无任何报错。
若你所在形态没有 skill 机制（只能用全局规则文件）：种子库留仓库 clone 目录（不得位于任何 workspace 内），用 clone 绝对路径引用。

核心原理：风格即基因，艺术家名即激活词

风格 = 色彩 × 光线 × 笔触 × 构图 × 情绪 五个基因位，调风格是调五个基因位，不是换个背景色
艺术家名字/风格名是 LLM 预训练美学基因的激活锚点——prompt 里直接点名「马远一角构图」「卡拉瓦乔 Tenebrism」比任何形容词都精准，因为预训练语料见过海量这些名字背后的作品与评论
本技能的职责：帮你选对激活词 + 把激活出的审美翻译成可执行参数

媒介翻译对照表（同一基因，五种落点）

选完风格基因后，按产出媒介翻译成对应参数：

媒介	基因落点
🖼️ 图片 / 生图	直接用种子的 AI Prompt Keywords + 艺术家激活词
🌐 HTML / 网页 / 长图	CSS 变量块 + 排版规则（`code-genes/` 有现成参数可抄）
📊 PPT / 幻灯	版式气质：底色、标题处理、每页信息密度、图表风格
📄 Word / 文档排版	字体气质（衬线/无衬线）+ 行距字距 + 标题层级样式 + 页眉页脚与色彩点缀；正式文档优先「清晰理性/庄重典雅」档
🏡 生活与社会关系	基因语言迁移到材质与场景：穿搭配色逻辑、家居氛围（光/材质/留白）、礼物与请柬的仪式感设计

审美咨询类任务（不产出文件，如"帮我配一身衣服"）同样走决策流程：先问/定情绪，再给带基因依据的方案。

决策流程（开工前必走，不许跳步）

情绪定位：先回答"读者该感受到什么？"（一个词，如敬畏/治愈/紧张/怀旧）
扫全景拿候选（不许只看首选表 —— 首选表只是入口之一，不是全部货架）：
- 机器选 → seedbrain match "<任务描述>" --profile profiles/taste.json 拿 IDF 加权的 top-K（跨层均衡，附命中信号与 idf）
- 人眼通读 → references/SEED-INDEX.md（全库等距平铺，每粒一行，无「首选/备选」层级差）
定 1 主 + 1 辅：从候选里挑匹配度最高的主基因 + 一个张力互补的辅基因
基因翻译：五基因 → 具体参数（CSS 变量 / AI Prompt Keywords / 版式气质）
一句话自检："这张作品的风格基因是什么（主+辅）？"答不上来 = 没想清楚，回到第 1 步
交付后记档：搭配与反馈写进 references/ART-SEEDS-LOG.md，并 seedbrain pick <id> --profile profiles/taste.json 记状态（已用过的种子在 14d 衰减窗口内会自然下沉，避免连续复读同一粒）

四层查询路由（按需加载，禁止全文注入上下文）

层	路径	何时查	内容
L0	本文件	每次视觉任务	决策流程 + 情绪索引表
L-B 全景索引	`references/SEED-INDEX.md` + `seedbrain`（引擎来自 `agent-seed-brain`）	每次视觉任务·选种前	全库等距平铺（374 粒）· IDF 加权 match · 状态记忆与时间衰减
L2a 情绪档	`references/moods/`（10 份）	知道目标情绪时	每情绪 2-3 粒主推种子五基因详解 + 媒介适配 + 翻车清单
L2b 代码基因卡	`references/code-genes/`（12 张）	写代码出图（网页/长图/ dashboard）	直接可抄的部署提示词 + CSS 变量块 + 反模式
L2c 风格章档	`references/chapters/`（24 章 + 后记）	要深挖某艺术家/流派细节时	《艺术风格百科大全》原文，每艺术家 2-3 句细述 + Prompt Keywords
L1 总库速查	`references/art-seeds.md`	按关键词全局搜种子时	全量种子的单行速查表

典型路径：情绪定位 → 查 moods/ 对应档案 → 若写代码再查 code-genes/ 对应基因卡抄参数 → 需要更多艺术家细节时进 chapters/。（下文速记 moods/xx、code-genes/xx 均相对 references/ 目录。）

常见信号快捷入口（入口之一，不是唯一货架）

⚠️ 下面这张表只是「高频信号 → 一份档案」的快捷跳转。它不是采样空间—— 真正的候选集来自 SEED-INDEX.md 全景（374 粒）或 seedbrain match。只在表内选种，会退化成「永远用最显眼那几粒」。

目标情绪	情绪档	首选种子
🌑 科技酷感	`moods/01`	赛博霓虹（`code-genes/02`）
✨ 梦幻治愈	`moods/02`	吉卜力手绘（`code-genes/05`）
🔥 热血冲击	`moods/03`	Trigger 爆裂色、鸟山明
🏛️ 庄重典雅	`moods/04`	新古典、北宋全景山水
🌸 甜美容爱	`moods/05`	洛可可粉彩（`code-genes/08`）
🌫️ 孤独诗意	`moods/06`	一角留白（`code-genes/03`）、Friedrich 雾景
🃏 荒诞幽默	`moods/07`	马格利特、哈林（`code-genes/09`）
💀 暗黑崇高	`moods/08`	浪漫主义废墟（`code-genes/12`）
📊 清晰理性	`moods/09`	瑞士平面（`code-genes/01`）
🏮 东方古韵	`moods/10`	宋韵双拼（`code-genes/03`）

（不在表中的情绪：先归到最接近的一档，再从 chapters/ 自由选种。）

铁律

先定情绪，再选基因，顺序不能反
不许只在首选表里选种子：优先跑 seedbrain match 或通读 SEED-INDEX.md，让全库 374 粒都有被选中的机会
一组作品一个主基因，最多加一个辅基因点睛（敢混搭，但只混一层）
颜色 ≤4；全图一种圆角 / 一种描边风格
留白与克制是普世高级感；东方基因（留白/水墨/线描/气口）是差异化武器
用户反馈 > 教条：好用的搭配标记复用，翻车的记录规避

红线

不许不经种子库直接输出"默认审美"（通用卡片/蓝紫渐变/AI 味配色）
不许把多种笔触/多种描边风格混在同一作品里
种子库按需查询，禁止把 chapters/ 或 art-seeds.md 全文读进上下文
种子库是活文档：用户点名的风格、验证过的新配方，随时补录进对应层 + 风味日志

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

and 90 more files.

Frequently asked questions

What does the Taste AI skill do?

Why use Taste on TypingMind?

Because you install it once and use it with any model. Taste is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Taste in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/hhx465453939/agent-taste-seed-system/tree/main. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Taste?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Taste?

As many as you like. As long as a model supports skills, you can use Taste with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Taste AI skill free?

It is published on GitHub by hhx465453939. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Openclaw Test Heap Leaks

openclaw

OrganizationPopular

Investigate OpenClaw pnpm test memory growth, Vitest OOMs, RSS spikes, and heap snapshot deltas.

Prototype Openclaw Tui

openclaw

OrganizationPopular

Build throwaway, fixture-driven OpenClaw Clack or Pi TUI prototypes and compare multiple interactive variants side by side in tmux without running the full application or touching live state.

Slacrawl

Slack archive: search, sync freshness, threads/DMs, SQL counts, and Slacrawl repo work.

Obsidian Vault Maintainer

openclaw

OrganizationPopular

Maintain an Obsidian-friendly memory wiki vault with wikilinks, frontmatter, and official Obsidian CLI awareness.

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

🎨 Taste — 审美操作系统（四层知识库）

种子库装在哪（引用规则）

核心原理：风格即基因，艺术家名即激活词

媒介翻译对照表（同一基因，五种落点）

决策流程（开工前必走，不许跳步）

四层查询路由（按需加载，禁止全文注入上下文）

常见信号快捷入口（入口之一，不是唯一货架）

铁律

红线

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now