skills.rest
gestaosian
gestaosianOFFICIAL
Explore all skills by this author →
GitHub

sian-dag-review

Reviews SIAN platform Airflow DAG files against the full factory checklist before commit.

Updated May 21, 2026
Search Tags:#code review#checklist#data engineering#airflow#BigQuery#iceberg#dag
One-click install
npx skills add https://github.com/gestaosian/gtsi-claude-skills --skill sian-dag-review-gestaosian
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sian-dag-review
Source: https://github.com/gestaosian/gtsi-claude-skills/tree/main/plugins/gtsi-ops-plugin/skills/sian-dag-review
Command: npx skills add https://github.com/gestaosian/gtsi-claude-skills --skill sian-dag-review-gestaosian

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Data engineers on the SIAN platform need to verify that Airflow DAG files follow platform standards (naming, factory usage, credential handling, Iceberg configuration) before committing or opening a PR, and manual review is error-prone. ## Core Features & Use Cases - Full checklist review: Validates file naming ({sistema}__{tabela}__{cliente}__{frequência}.py), dag_id matching, directory location, sys.path depth, factory usage, Iceberg cluster_by, credential handling via Airflow Variables, and pause-by-default settings. - Structured verdict: Emits an approve / approve-with-reservations / reject verdict following the gtsi-code-review pattern, shown inline and saved to .gtsi/reviews/<author-slug>/. - Cross-skill context: References related sian-* skills (dag-factory, iceberg-setup, new-system-checklist, data-destination, data-correction-policy) for detailed guidance on each checklist item. - Use Case: Before committing dags/protheus/financeiro/protheus__se1__sian__daily.py, ask for a review and receive a checklist-based verdict with a saved report. ## Quick Start Ask the assistant to review the DAG file at dags/protheus/financeiro/protheus__se1__sian__daily.py before committing it.

Frequently Asked Questions about sian-dag-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review an Airflow DAG before committing it?▼

Provide the DAG file path or diff and the checklist is applied covering naming, location, sys.path depth, factory usage, credentials, and scheduling. You receive an inline verdict plus a saved report in .gtsi/reviews/<author-slug>/.

What naming convention does the SIAN DAG checklist enforce?▼

Files must follow {sistema}__{tabela}__{cliente}__{frequência}.py with lowercase system and table names, double-underscore separators, daily or intraday frequency, and a dag_id exactly matching the filename without .py.

When should a SIAN DAG use Iceberg table format?▼

Use table_format="iceberg" for large tables or continuous growth, with cluster_by set to up to four frequently filtered columns. A hard delete handler is required if the source physically deletes records.

Can a DAG be approved without using the system factory?▼

No. Systems with an existing factory (Protheus, RM, Pontotel, TDMAX, Supabase) must use it, and manually built DAGs for those systems are rejected. Hardcoded credentials and unpaused-by-default DAGs are also rejected.

How many dirname() calls should sys.path use in SIAN DAGs?▼

DAGs located at dags/<sistema>/<domínio>/ (three levels below root) need three dirname() calls, while DAGs at dags/<sistema>/ (two levels) need two. This ensures the factory modules resolve correctly.

Related Skills

Explore More Data & Analytics Skills→
ivanshamaevivanshamaev
14

airflow_dag_factory

Generate Apache Airflow DAGs declaratively from YAML using dag-factory.

Community
Advanced
datasurfacedatasurface

Start Initial Ingestion

Tag DataSurface Yellow model releases, provision Kubernetes secrets, and trigger Airflow ingestion DAGs.

Official
Advanced
uk-agentsuk-agents
9

policy-starter

Draft a UK internal AI usage policy from published model guidance.

Official
Advanced
nikegeorgian-stacknikegeorgian-stack

fst-design-system

Documents the as-built Tailwind v4 design system of the FiberCell Otgruzka web UI.

Community
Intermediate