convez.ai
SkillProdutividadeIntermediário
andy-builds-ai

Claude Code Skills para Python

andy-builds-ai/claude-code-skills

Skills automáticas para código Python mais estruturado e debugging melhor

Ver no GitHub
InicialPluginScore · de 0 a 100

Conjunto de skills que ativam automaticamente durante o desenvolvimento Python com Claude Code. Inclui orientações para escrita de código mais limpo (uma responsabilidade por função), scaffolding de projetos novos e explicação contextual de erros em tracebacks.

Instalação

$npx convez-ai andy-builds-ai/claude-code-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
0
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 4 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Refatorar funções com múltiplas responsabilidades em componentes focados
  • Scaffoldear novos projetos Python (módulos, CLIs, APIs) com estrutura padrão
  • Entender e diagnosticar erros Python através de análise de tracebacks

Veredito

Ferramenta sólida que automatiza decisões arquiteturais comuns em Python, economizando tempo na estruturação de código e debug. Ideal para desenvolvedores que querem manter padrões sem pensar manualmente em cada função.

README

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

Claude Code Skills

A small collection of Claude Code skills for Python development.

What is a skill?

A skill is a single SKILL.md file that Claude Code activates automatically when a certain situation arises. You don't call it by hand — the context triggers it. Each file has a short description in its frontmatter that tells Claude Code when the skill is relevant. When the current work matches that description, the skill's instructions are loaded and shape how Claude Code answers.

In other words: you keep working normally, and the right skill switches itself on at the right moment.

To use them, copy the skill folders into the ~/.claude/skills/ directory (or a project-level .claude/skills/).


bau-begleiter

What it does. Guides active building: one function does exactly one thing, code grows block by block (write the function, then call it right away with a real value), and assumptions about the inputs are checked at the start of each function.

When it triggers. As soon as code is actively being written — functions, small modules, scripts. Not during design (before the code), not during the final review (after the code). The trigger is the moment code is being written.

Example. You sketch a single function that does several things at once:

def save_user(data):
    # validates the data AND writes it to the database AND sends a welcome mail
    ...

The skill catches the "and": three responsibilities in one function. It splits them into validate_user(data), persist_user(data), and notify_user(user) — each with one job, each callable on its own.


module-scaffolder

What it does. Sets up the skeleton of a new project and adapts to the project type (Python module, CLI tool, API service, or learning script). It provides the directory structure, a .gitignore, a README stub, and a requirements.txt base.

When it triggers. When a new project starts — a new project folder is created or you say "I'm starting a new project". It runs before the first line of code, not after.

Example. You say: "I'm starting a new Python module, it should become an HTTP client wrapper." The skill sets up the repo folder http-client-wrapper/ with the inner module http_client_wrapper/, writes the .gitignore first (with .env in it), adds an English README with the standard sections, drops in the logging pattern, and makes the first commit feat: initial scaffold.


error-explainer

What it does. Walks you through a Python traceback: read it from the bottom up, understand the mechanics behind the error first, and only then look for the fix. It works in a learning mode (guides you to the cause) and a tool mode (direct diagnosis).

When it triggers. As soon as a Python traceback appears in the output. The trigger is binary: is there a traceback, yes or no. It does not trigger when code runs but returns the wrong result — only when Python actually throws an error.

Example. This shows up in the output:

Traceback (most recent call last):
  File "example.py", line 12, in <module>
    print(numbers[5])
IndexError: list index out of range

The skill points at the bottom line first ("the error is IndexError, not the file path"), explains what Python checks on a list access, and asks a diagnostic question instead of handing over the fix: "Run print(len(numbers)) right before line 12 — how many elements does the list actually have at that moment?"

Ver o original no GitHub

Bom para

Relacionados