Not in the top 1,000 by GitHub stars

qybaihe/hash-design

Let a true-random hash pick the design direction, not the model. MCP server + Agent Skill + CLI + TypeScript library mapping hashes onto 40 real-world design genres.

History

Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.

View weekly changes

skills/hash-design/SKILL.md

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Instructions only

First observation → 663fe4d1b26c

No earlier version was observed for comparison.

Compare on GitHub

View on GitHubAdd to my skills

Verified commit 663fe4d1b26cdce0ee6cfeb2be98a523e06a317e

4GitHub stars
Not availablestars in 7 days
1skill in this repository
117commits in 30 days

Heat

Insufficient history

Not enough history yet to measure heat. It appears after a few days of data.

Heat never changes the order by stars, and sponsorship never touches it. How heat works

Stars over time

Nov 3, 2026Nov 4, 20264

30-day change not yet available: not enough history.

Individual skills in this repository

How it works

hash-design

When your agent uses it

Use when building or redesigning any web page, landing page, one-page site, slide deck, mobile app screen, dashboard, poster / social card or HTML email where the result must have real design character and must NOT look AI-generated (no Inter, purple gradients, centered hero with two buttons, three identical cards). Rolls true-random hashes through the hash-design MCP server (each hash is one design direction: its bytes set the palette, faces, grid, motif, motion, voice and twist directly, with no style library), freezes N design briefs (default 5, up to 12; scenario web, deck, app, dashboard, poster or email), replies with one line per direction and waits; on request renders independent HTML previews plus a side-by-side compare page; then builds the picked one and self-critiques it. Triggers - "做个页面/官网/落地页/一页站", "做个 PPT/幻灯片", "App 界面", "数据看板", "海报", "去 AI 味", "要有设计感", "landing page", "slide deck", "make it not look like AI", "give me design options".

What it loads into context

Frontmatter

252 tokens

Always available to the agent

SKILL.md

3,427 tokens

When the skill is activated

Reference documents

0 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

  1. 何时使用
  2. 前置条件
  3. 工作流(严格按顺序)
  4. 硬性规则
  5. MCP 工具 ↔ CLI 对照
  6. 目录约定

Files

  • SKILL.mdcore11,912 bytes
  • SKILL.mdcore11,912 bytes
Read SKILL.md

# hash-design:用外部随机打破默认审美

LLM 做页面会塌缩到众数审美。这个 Skill 的做法是:**风格不由你选**。先用真随机哈希掷骰:**哈希本身就是设计方向**,它的每个字节按固定规则直接决定一个参数(配色的色相与明度、标题和正文字体、栅格、首屏锚点、母题、材质、动效、文案语气、变数),没有任何风格库可选。N 个哈希(默认 5 个)就是 N 份 brief,你只负责**忠实执行** brief,把选择权交还给用户。

## 何时使用

- 用户要做/重做一个网页、落地页、一页站、活动页、作品集首页、UI 视觉稿;或幻灯片/PPT、手机 App 界面、数据看板、海报/社交卡片、HTML 邮件。
- 用户说"要有设计感""别像 AI 做的""给我几个方向"。
- 不适用:纯后端、已有严格设计系统且用户要求遵守的改动(此时只用 `critique_page` 自检即可)。

## 前置条件

- 已连接 MCP server `hash-design`(工具:`roll_designs`、`get_brief`、`critique_page`、`build_compare_page`)。
- 没有 MCP 时用 CLI,参数相同(见文末对照表):`npx -y hash-design roll …`、`npx -y hash-design critique …`、`npx -y hash-design compare …`;在源码目录里是 `node packages/core/dist/bin.js …`。`skill install` 安装后会打印当前可用的 CLI 和 MCP 命令。
View the complete source on GitHub

Skills included

  • hash-design

Repository facts

License
MIT
Created
Nov 1, 2026
Last pushed
Nov 3, 2026
Forks
1
Open issues
0
Project activity
Active

Before you install

Instructions only

Automated signals, not a guarantee

Checked Nov 6, 2026

Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.

Show 1 note
  • skills/hash-design/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • No broad tool permissions
  • No dangerous shell commands
  • No network access
  • No hidden or obfuscated text
  • No instruction overrides
  • No executable scripts
  • No access to secrets
  • No requests to vote in Hot

How the signals fit together

Facts from a static read of the files at the pinned commit. They describe what the code does, not why.

No combined signals at this commit.

Works with

Agent compatibility, from the files in this repository
AgentStatusEvidenceInstall to
Claude CodeCompatibleStandard SKILL.md skills/hash-design/SKILL.md.claude/skills/<skill-name>/
CodexCompatibleStandard SKILL.md skills/hash-design/SKILL.md.agents/skills/<skill-name>/
CursorCompatibleStandard SKILL.md skills/hash-design/SKILL.md.cursor/skills/<skill-name>/
Gemini CLICompatibleStandard SKILL.md skills/hash-design/SKILL.md.gemini/skills/<skill-name>/
GitHub CopilotCompatibleStandard SKILL.md skills/hash-design/SKILL.md.github/skills/<skill-name>/
OpenCodeCompatibleStandard SKILL.md skills/hash-design/SKILL.md.opencode/skills/<skill-name>/
  • Native The repository ships a manifest, folder or frontmatter field for this agent.
  • Compatible A standard SKILL.md that this agent can read.
  • Unknown No evidence either way.

Derived automatically from paths, manifests and frontmatter. Not tested on each agent.