python-foundational

Práticas fundamentais de Python, expressões idiomáticas e fundamentos de qualidade de código - Fornecido por microsoft/hve-core

npx skills add https://github.com/microsoft/hve-core --skill python-foundational

Python Foundational Coding Standards Skill

Overview

Foundational Python excellence that every diff must satisfy. This skill is loaded first for any .py change. All higher-order skills build on it.

This content is a skill rather than an instructions file for three reasons: skills are distributed through the CLI plugin and VS Code extension without requiring consumers to copy files into their repo; new language skills can be added without modifying the review agent itself; and skills are loaded on demand, keeping the context window small when the diff contains no Python.

Core Checklist

1. Readability & Style

  • Use Python naming: PascalCase classes, snake_case functions/variables, UPPER_SNAKE_CASE constants, _ private members.
  • Group imports: stdlib → third-party → local (blank line between groups, no trailing whitespace).

2. Pythonic Idioms

  • Prefer comprehensions for simple transforms; use explicit loops for complex logic/side-effects.
  • Always use with for files, locks, DB connections.
  • Prefer dataclass / NamedTuple / Enum for data holders.
  • Use pathlib over os.path; timezone-aware datetime when relevant.
  • Use * keyword-only arguments for multi-optional functions.
  • Never use mutable defaults or global/nonlocal unless strictly required.

3. Function & Class Design

  • Keep functions small and single-responsibility.
  • Add docstrings to all public APIs (follow repo style).
  • Document unavoidable side-effects.
  • Follow codebase’s class-member ordering (if defined).

4. Type Safety Foundations

  • Add type hints to all public APIs, module vars, and class attributes.
  • Use PEP 695 (3.12+) or TypeVar for generics.
  • Avoid Any except in thin wrappers.

5. Error Handling

  • Raise specific exceptions; never bare except: (broad except Exception: only at app boundaries with logging).
  • No silent failures or generic error messages.
  • Provide context, expected state, and guidance in every exception.

6. Anti-Patterns to Avoid

  • Never use eval, exec, or pickle on untrusted data.
  • Never hard-code secrets.

7. Maintainability

  • Prefer self-documenting code; comments only for "why".
  • Use structured logging instead of print.
  • Flag overly long/complex functions that resist testing.

8. Architectural Fit

  • Align with existing patterns; do not re-implement shared functionality or bypass established layers.
  • Place code in the correct module/package.

9. Design Principles

  • Eliminate duplication: extract repeated logic into a shared helper so fixes propagate automatically.
  • Prefer the simplest implementation that satisfies current requirements. Introduce abstractions only when a second concrete use case appears.
  • Limit change breadth: every modified line should trace to the stated purpose of the change.
  • Before flagging seemingly unused code, verify it is not a protocol implementation, framework hook, public API, or entry point invoked externally.
  • Match solution complexity to problem complexity. A duplicated function warrants a shared helper, not an event-driven architecture.

References

FileCoversPurpose
design-principles.mdSection 9Rationale and examples for the design principles
code-style-patterns.mdSections 1–5Concrete code examples for style, idioms, type safety, class design, and error handling

Severity Rubric

SeverityDefinition
HighCauses incorrect behavior, data loss, or security exposure at runtime
MediumDegrades maintainability, readability, or violates a project convention with no immediate runtime impact
LowCosmetic, stylistic, or minor improvement opportunity

Troubleshooting

SymptomCheck
Skill not loadedConfirm the diff contains .py files. The agent selects skills by matching file types in the changed files against skill descriptions.
No findings generatedVerify the Skills Loaded footer in the review output lists python-foundational. If listed but no findings appear, the diff may already satisfy the checklist.
Severity seems miscalibratedCompare against the Severity Rubric above. High requires runtime impact; medium is maintainability-only.

Contributing

Follow these conventions when extending this skill:

  • Checklist items belong in SKILL.md. Each bullet is a single, actionable check an agent or reviewer can apply to a diff. Keep bullets concise: one sentence, no code.
  • Reference files live in references/ and provide examples or rationale for the covered checklist items. Each reference file covers a contiguous range of sections. Update the References table when adding a new file.
  • Before adding a new checklist item, confirm it does not duplicate an existing bullet in any section. Place it in the section that matches its primary concern. If it spans concerns, prefer the more specific section.
  • Name new reference files after the topic they cover (e.g., async-patterns.md). Include a frontmatter description that states which sections the file supports. Add a row to the References table in SKILL.md.
  • Checklist items and examples must be portable across codebases. Use generic module names and describe anti-patterns by their behavior, not by specific directory or file names.

Mais skills de microsoft

oss-growth
microsoft
Persona de growth hacker OSS
official
microsoft-foundry
microsoft
Implantar, avaliar e gerenciar agentes Foundry de ponta a ponta: build Docker, push ACR, criação de agente hospedado/prompt, inicialização de contêiner, avaliação em lote, avaliação contínua, fluxos de trabalho do otimizador de prompt, agent.yaml, curadoria de conjunto de dados a partir de rastros. USE PARA: implantar agente no Foundry, agente hospedado, criar agente, invocar agente, avaliar agente, executar avaliação em lote, avaliação contínua, monitoramento contínuo, status da avaliação contínua, otimizar prompt, melhorar prompt, otimizador de prompt, otimizar instruções do agente, melhorar agente...
officialdevelopmentdevops
azure-ai
microsoft
Use para Azure AI: Search, Speech, OpenAI, Document Intelligence. Ajuda com pesquisa, busca vetorial/híbrida, fala para texto, texto para fala, transcrição, OCR. QUANDO: AI Search, pesquisa de consulta, busca vetorial, busca híbrida, busca semântica, fala para texto, texto para fala, transcrever, OCR, converter texto em fala.
officialdevelopmentapi
azure-deploy
microsoft
Execute implantações do Azure para aplicativos JÁ PREPARADOS que possuem arquivos .azure/deployment-plan.md e de infraestrutura existentes. NÃO use esta skill quando o usuário pedir para CRIAR um novo aplicativo — use azure-prepare. Esta skill executa comandos azd up, azd deploy, terraform apply e az deployment com recuperação de erros integrada. Requer .azure/deployment-plan.md do azure-prepare e status validado do azure-validate. QUANDO: "executar azd up", "executar azd deploy", "executar implantação",...
officialdevopsaws
azure-storage
microsoft
Serviços de Armazenamento do Azure, incluindo Blob Storage, File Shares, Queue Storage, Table Storage e Data Lake. Responde a perguntas sobre camadas de acesso ao armazenamento (hot, cool, cold, archive), quando usar cada camada e comparação entre elas. Oferece armazenamento de objetos, compartilhamentos de arquivos SMB, mensagens assíncronas, NoSQL chave-valor e análise de big data. Inclui gerenciamento de ciclo de vida. USE PARA: blob storage, file shares, queue storage, table storage, data lake, upload de arquivos, download de blobs, contas de armazenamento, camadas de acesso,...
officialdevelopmentdatabase
azure-diagnostics
microsoft
Depure problemas de produção no Azure usando AppLens, Azure Monitor, integridade de recursos e triagem segura. QUANDO: depurar problemas de produção, solucionar problemas do Serviço de Aplicativo, alto uso de CPU no Serviço de Aplicativo, falha de implantação do Serviço de Aplicativo, solucionar problemas de aplicativos em contêineres, solucionar problemas de funções, solucionar problemas do AKS, kubectl não consegue conectar, falhas do kube-system/CoreDNS, pod pendente, crashloop, nó não pronto, falhas de atualização, analisar logs, KQL, insights, falhas ao puxar imagem, problemas de inicialização a frio, falhas de sonda de integridade,...
officialdevopsdevelopment
azure-prepare
microsoft
Prepare aplicativos do Azure para implantação (infra Bicep/Terraform, azure.yaml, Dockerfiles). Use para criar/modernizar ou criar+implantar; não para migração entre nuvens (use azure-cloud-migrate). NÃO USE PARA: aplicativos copilot-sdk (use azure-hosted-copilot-sdk). QUANDO: "criar aplicativo", "construir aplicativo web", "criar API", "criar API HTTP serverless", "criar frontend", "criar backend", "construir um serviço", "modernizar aplicativo", "atualizar aplicativo", "adicionar autenticação", "adicionar cache", "hospedar no Azure", "criar e...
officialdevelopmentdevops
azure-validate
microsoft
Validação pré-implantação para prontidão do Azure. Execute verificações aprofundadas de configuração, infraestrutura (Bicep ou Terraform), atribuições de função RBAC, permissões de identidade gerenciada e pré-requisitos antes de implantar. QUANDO: validar meu aplicativo, verificar prontidão para implantação, executar verificações de pré-voo, verificar configuração, verificar se está pronto para implantar, validar azure.yaml, validar Bicep, testar antes de implantar, solucionar erros de implantação, validar Azure Functions, validar function app, validar serverless...
officialdevopstesting