DES-SKILL

Use when designing the global data architecture, including storage, compute, integration patterns, layer strategy, environment promotion, security posture, observability, governance, cost, operational burden, and decision reversibility.

Use when creating the Bronze Layer Specification for raw/near-raw storage, source fidelity, ingestion metadata, schema drift, replay, quarantine, retention, access control, and lineage before Silver design.

Use when designing CI/CD, testing strategy, release gates, deployment promotion, rollback, release evidence, and environment validation for data products.

Use when designing cost, performance, scalability, storage, compute, query, serving, monitoring, and FinOps optimization strategies for data products.

Use when creating the Data Contract Specification for source feeds, Silver datasets, Gold datasets, serving outputs, dashboards, semantic models, APIs, ML/AI datasets, external exports, or downstream systems that need stable structure, meaning, freshness, quality, access, ownership, lineage, and change behavior.

Use when defining or designing data quality dimensions, rules, thresholds, severity, gates, failure handling, ownership, and validation expectations across Bronze, Silver, Gold, contracts, transformations, and serving outputs.

Use when creating the Gold Layer Specification for curated, consumer-ready marts, metric-ready datasets, reporting outputs, feature datasets, API/read-model outputs, or product-aligned datasets derived from trusted Silver data.

Use when creating the Ingestion Specification for any data engineering project, including source-specific ingestion pattern, frequency, access method, checkpointing, idempotency, replay, schema drift, security, error handling, landing expectations, and ingestion observability.

Use when designing data lineage, catalog metadata, business/technical/operational/reference metadata, run evidence, ownership, trust metadata, and auditability for data products.

Use when designing schedules, triggers, dependencies, quality gates, retries, failure handling, recovery, backfills, run evidence, alerting, incident behavior, and operational monitoring for data pipelines.

Use when creating the Project Evaluation Report to evaluate business value, technical readiness, release readiness, adoption evidence, risks, lessons learned, and final closeout for data products.

Use when designing semantic models, metrics layers, measures, dimensions, hierarchies, relationships, trust status, freshness/quality visibility, and business-friendly data consumption.

Use when designing how data products are served to consumers via dashboards, BI datasets, semantic models, APIs, ML/AI datasets, reverse ETL, exports, data sharing, embedded analytics, or AI/data agents.

Use when creating the Silver Layer Specification for cleaned, normalized, conformed, validated, deduplicated, keyed, traceable, domain-aligned entity/event datasets before Gold design.

Use when designing transformation logic, dependency flow, model boundaries, incremental behavior, business rules, conformance rules, metric logic, validation expectations, and contract-aligned outputs across Bronze, Silver, Gold, and serving-ready layers.

Use when implementing, reviewing, or planning dbt work inside a DES workflow, including model structure, sources, staging, intermediate models, marts, tests, unit tests, docs, contracts, selectors, local dbt-duckdb execution, CI checks, and dbt MCP usage.

Use when designing, implementing, testing, debugging, deploying, or reviewing Apache Airflow orchestration inside a DES workflow, including Dags, task boundaries, schedules, retries, backfills, data quality gates, dlt/dbt/Provero/Fabric integration, lineage, observability, alerts, MCP usage, and operational evidence.

Use when designing, implementing, reviewing, or validating dlt ingestion pipelines inside a DES workflow, including REST APIs, SQL databases, filesystems, cloud storage, sources, resources, write dispositions, incremental state, schema evolution, schema contracts, destinations, Bronze loading, and evidence handoff.

Use when using DuckDB as a local-first execution engine inside a DES workflow for reading files, profiling datasets, validating schema and grain, prototyping SQL transformations, converting files, creating local evidence, or preparing handoff to dbt, dlt, Fabric, Provero, or other data engineering stacks.

Use when designing, implementing, reviewing, or collecting evidence from Provero data quality checks inside a DES workflow, including source profiling, Bronze/Silver/Gold validation, data contracts, SLA checks, anomaly detection, result history, reports, alerts, dbt interoperability, and Airflow integration.

Use when identifying, probing, profiling, comparing, or approving source systems before domain modeling, ingestion, architecture, modeling, or transformation design.

Use when starting a data engineering project, the project idea is vague or solution-first, business context is missing, or the workflow router selects Phase 1.

Use when Phase 1 context and handoff exist, and business questions are vague, duplicated, unprioritized, unmapped to consumers, or needed before KPI and data product design.

Use when requirements, KPIs, business questions, dashboards, APIs, semantic models, ML datasets, or data outputs need a clear data product boundary, consumer, owner, lifecycle, trust level, or service expectation.

Use when identifying core business concepts, glossary, ontology-lite concepts, entities, events, relationships, grains, terminology, ownership, source alignment, and ambiguity risks before modeling tables.

Use when designing governance policy, data classification, access control, privacy handling, masking, encryption, retention, audit, approval workflow, and security risk controls for data products.

Use when Phase 2 business questions need to become measurable data engineering requirements, KPI definitions, SLA expectations, acceptance criteria, or metric ownership decisions.

Use when the user wants to test understanding of a DES artifact, generate concept-check questions, or turn learning notes into a quiz.

Use when the user wants to learn from a DES artifact, understand the Data Engineering concepts behind an artifact, or generate learning notes for a completed DES phase.

Use when the user wants personalized feedback on their understanding of a DES artifact, quiz answer, explanation, or Data Engineering concept.

Use when the user wants a structured learning path across DES phases, wants to study Data Engineering through DES artifacts, or wants a personalized plan based on completed artifacts.

Use when the user wants a summary review of learning progress across DES phases, weak concepts, readiness, and next study actions.

Use when learning notes, quizzes, gap reports, or Socratic coaching sessions have changed and the DES learning status file should be updated.

Use when the user should learn a DES concept through guided questions rather than direct explanation.

Use when the user needs help choosing the right DES skill, understanding the DES workflow, finding the next step, or routing between lifecycle, support, and learning skills.

Use when starting, resuming, auditing, or coordinating any data engineering project with DES-SKILL installed.

Use when defining ownership, stewardship, access control, privacy, lineage, retention, compliance, and change management for data products.

Use when defining data quality dimensions, thresholds, anomaly baselines, orchestration, observability, retries, alerts, or operational response rules.

Use when reviewing lineage, metadata, catalog stewardship, schema registry compatibility, privacy, PII, masking, tokenization, RBAC, RLS, audit, retention, or security policy.

Use when designing semantic models, serving layers, certified metrics, BI access, query performance, caching, RLS, APIs, reverse ETL, or consumer-facing datasets.