菜单
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".
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
| name | schema-audit |
| description | 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". |
| context | fork |
| agent | Explore |
Pesquise e audite todos os schemas Pydantic do repositório.
schemas/ com Glob.py encontradoetl-monorepo/jobs/ e leia cada main.pyetl-monorepo/jobs/job_template/main.py como baselineBaseModel em todo o repositório para encontrar schemas inlinereferences/bigquery-pydantic-types.md para mapeamento de tiposPara cada schema encontrado, verifique:
Optional[T] ou T | None?datetime em vez de str?Decimal em vez de float?field_validator para regras de negócio?min_length ou pattern?Literal ou Enum?schemas/ é importado corretamente no job?main.py que deveriam ser extraídos?REQUIRED estão sem Optional no Pydantic?# Schema Audit Report
## Resumo
- Schemas analisados: N
- Issues encontradas: N (X critical, Y warning, Z info)
## Por schema
### schemas/exemplo.py
- [CRITICAL/WARNING/INFO] descrição
## Schemas inline (extrair para schemas/)
- caminho:linha — classe encontrada
## Campos inconsistentes entre schemas
- campo_a (schema_x.py) vs campo_b (schema_y.py)
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.
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".
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.