convez.ai
Skill
zenthos-z

my-skills

zenthos-z/my-skills

Ver no GitHub
InicialPluginScore · de 0 a 100

Claude Code skills monorepo

Instalação

$npx convez-ai zenthos-z/my-skills

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
0
Acompanhando
0
Saúde
28%
Último commit
há 2 semanas
Versão
—
Issues abertas
0
Contribuidores
1

README

A tradução para o português deste README ainda está na fila.


TL;DR

Problem: Claude Code is powerful, but repetitive tasks (diagram generation, report writing, image creation, structured analysis) require manual prompting every time.

Solution: Installable skill packs that teach Claude domain-specific workflows — validated syntax, structured outputs, and multi-step automation built in.

SkillOne-lineRuntime
mermaid-proMermaid diagrams with syntax validation + image exportNode.js
systems-thinkingStructured systems thinking interviews with session persistenceNone
qunribaoWeChat group daily/weekly report generation from chat dataPython
quick-imgFast image generation via DMX API (Gemini Flash)Python

Quick Start

# Install a single skill
npx skills add zenthos-z/my-skills/mermaid-pro

# Or install all skills
npx skills add zenthos-z/my-skills/mermaid-pro
npx skills add zenthos-z/my-skills/systems-thinking
npx skills add zenthos-z/my-skills/qunribao
npx skills add zenthos-z/my-skills/quick-img

After installation, skills activate automatically when you trigger their keywords in Claude Code (e.g., "generate a mermaid diagram", "systems thinking interview", "群日报").

Skills Overview

mermaid-pro

Professional Mermaid diagram generation with consistent styling.

  • 9-color semantic palette — color-coded by meaning (green=start, red=decision, blue=process)
  • Built-in syntax validation — every diagram is parsed and verified before output
  • Batch image export — convert all Mermaid blocks in a Markdown file to SVG/PNG
  • 7 diagram types — Flowchart, Sequence, Class, ERD, C4, State, Mindmap
  • 3 style presets — minimal, professional (default), colorful
  • Fully offline — no external API calls, runs locally with Puppeteer
# Setup after install
cd ~/.claude/skills/mermaid-pro/scripts && npm install

See mermaid-pro/README.md for detailed configuration.


systems-thinking

Structured systems thinking coach based on Dennis Sherwood's methodology.

  • 5-stage interview protocol — Claude guides you through problem boundary exploration, causal mapping, loop identification, diagram drawing, and leverage point discovery
  • Session persistence — progress auto-saved at each stage, resume across conversations
  • 15 thinking trap recognitions — catches linear thinking, symptom fixing, ignoring time delays, and more
  • System loop diagram syntax — text-based [A] --S--> [B] notation for causal links
  • Zero dependencies — pure prompt engineering, no runtime needed

No setup required. Just invoke and start the interview.


qunribao

WeChat group daily/weekly report generation system.

  • JSON-first architecture — structured JSON → validated with JSON Schema → Markdown tables
  • Memory mechanism — multi-version topic tracking with auto-cleanup (keeps last 10)
  • Describe mode — Vision API analyzes chat images instead of unreliable file paths
  • Feishu Bitable upload — direct JSON-to-Bitable payload construction
  • Privacy scanner — pre-commit hook detecting group IDs, API keys, phone numbers
  • Graceful degradation — works without optional quick-img dependency
# Requires Python 3.8+ and a running WeFlow instance (port 5031)
# First-time setup: run the init wizard in Claude Code

See qunribao/README.md for WeFlow configuration and data pipeline details.


quick-img

Fast image generation via DMX API (Gemini 3.1 Flash Image Preview).

  • 3 usage modes — Direct Prompt, Template (Markdown → infographic), Refine (condense → confirm → generate)
  • 14 aspect ratios — social media (4:5), presentations (16:9), mobile (9:16), and more
  • 4 resolution tiers — 0.5K quick preview to 4K print quality
  • Batch "gacha" mode — same prompt, multiple distinct results via API randomness
  • SiYuan integration — read SiYuan notes as image generation prompts
  • Template system — custom .txt templates with {{content}} placeholder
# Requires Python 3.8+ and a DMX API key
# Set your key in assets/.env: DMX_API_KEY=sk-...

See quick-img/README.md for API key setup and template customization.

Cross-skill Dependencies

graph LR
    qunribao -->|optional cover image| quick-img
SkillDepends OnRequired?
qunribaoquick-imgNo — gracefully degrades without it

All other skills are fully independent.

Installation

Method 1: npx skills add (Recommended)

npx skills add zenthos-z/my-skills/<skill-name>

Method 2: Manual copy

# Copy a single skill
cp -r mermaid-pro ~/.claude/skills/

# Copy all skills
cp -r mermaid-pro systems-thinking qunribao quick-img ~/.claude/skills/

Method 3: Clone the repo

git clone https://github.com/zenthos-z/my-skills.git
# Then copy desired skills to ~/.claude/skills/

Requirements Summary

SkillRuntimeExternal Services
mermaid-proNode.js >= 18None (fully offline)
systems-thinkingNoneNone
qunribaoPython 3.8+WeFlow (local, port 5031)
quick-imgPython 3.8+DMX API (cloud)

Contributing

Contributions are welcome. Each skill is self-contained — feel free to add new skills or improve existing ones.

License

All skills are released under the MIT License. See individual skill directories for details.

Ver o original no GitHub

Bom para

Relacionados