菜单
Reviews all ETL jobs for structured logging compliance and observability best practices. Use when the user mentions "logs", "logging", "observability", "monitoring", "structured logging", "log review", or "check if jobs are logging correctly".
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when the user has rough meeting notes, transcripts, or agendas and wants them converted into clean notes under 4 - Meetings with clear decisions, actions, stakeholders, and project links.
Use when the user wants to clean up or expand project notes, connect them to stakeholders and references, and extract reusable knowledge into 3 - Knowledge without losing project-specific context.
Use when the user wants to sort new material in this vault, especially items in 0 - Staging or 6 - Notes, and needs help deciding what should become a stakeholder, project, knowledge, meeting, or reference note.
Audits all Pydantic schemas in the ETL monorepo for consistency, missing validations, type safety issues, and alignment with BigQuery tables. Use when the user mentions "audit schemas", "check schemas", "schema review", "Pydantic review", "data quality", or "validation gaps".
Scaffolds a complete ETL job for the monorepo with Pydantic schema, structured logging, retry logic, DLQ, Terraform resource, and tests. Use when the user says "create job", "new job", "add pipeline", "ingest data", "scaffold job", or mentions creating a new data source extraction.
Reviews Terraform infrastructure code for security issues, best practices, and compliance with project standards. Use when the user mentions "review terraform", "check infra", "IaC review", "security review terraform", "infra audit", or asks about Terraform configurations in the project.
| name | observability-review |
| description | Reviews all ETL jobs for structured logging compliance and observability best practices. Use when the user mentions "logs", "logging", "observability", "monitoring", "structured logging", "log review", or "check if jobs are logging correctly". |
| context | fork |
| agent | Explore |
Analise todos os jobs ETL para garantir conformidade com o padrão de logging.
Consulte a referência completa em references/logging-spec.md.
Todo job DEVE emitir um log JSON no final da execução com os campos: event, job, timestamp, valid_records, dlq_records, duration_seconds, success.
etl-monorepo/jobs/main.pyprint(json.dumps e logging em todos os jobsobservability/ para configurações adicionaisevent — presente e descritivo?job — presente e match com o nome do diretório?timestamp — formato ISO 8601 com Z?valid_records — contagem de registros válidos?dlq_records — contagem de registros na DLQ?duration_seconds — tempo de execução?success — booleano derivado do resultado?json.dumps() em vez de f-string para logs?print() para mensagens que não são logs estruturados?# Observability Review
## Compliance Score: X/Y jobs conformes
## Por job
### jobs/{nome}
- Status: COMPLIANT / NON-COMPLIANT
- Campos faltando: [lista]
- Issues: [lista com severidade]
## Recomendações globais
- [melhorias que se aplicam a todos os jobs]