starlake-ai/starflow-help

Starflow Help & Navigation

Overview

Helps users navigate the Starflow framework by analyzing the current project state, available artifacts, and recommending the appropriate next workflow or agent.

Role Guidance: Act as a helpful guide who understands the full Starflow framework and Starlake ecosystem.

Starflow Workflow Phases

Phase 1: Discovery

Start here when beginning a new data project.

• starflow-domain-discovery — Map data domains, sources, and ownership
• starflow-source-analysis — Deep-dive into a specific data source

Phase 2: Architecture

After discovery, design the data platform.

• starflow-create-data-architecture — Design layers, engines, governance
• starflow-schema-design — Define Starlake-compatible table schemas

Phase 3: Pipeline Design

With architecture in place, design the pipelines.

• starflow-create-pipeline-spec — Full ETL/ELT pipeline specification
• starflow-transform-design — Design SQL transformations
• starflow-orchestration-design — Design scheduling and DAGs

Phase 4: Implementation

Build and deploy the pipelines.

• starflow-dev-pipeline — Implement pipeline from spec
• starflow-sprint-planning — Plan and track sprint progress
• starflow-code-review — Review pipeline configuration and SQL

Available Agents

| Agent | Name | Specialty | |-------|------|-----------| | data-analyst | Lea | Domain discovery, source analysis | | data-architect | Winston | Architecture, schemas, pipeline design | | data-engineer | Amelia | Pipeline implementation, transforms | | data-quality-engineer | Quinn | Data quality, lineage, expectations | | platform-engineer | Max | Orchestration, deployment, operations |

Starlake Skills Integration

Starflow methodology skills work alongside the Starlake CLI skills. When executing workflow steps, use the relevant Starlake skills for implementation details:

| Starflow Phase | Starlake Skills to Use | |----------------|----------------------| | Schema Design | config (types reference), load (write strategies), infer-schema (from files) | | Pipeline Implementation | load, transform, extract-schema, extract-data | | Data Quality | expectations (Jinja2 macros, validation syntax) | | Lineage Review | lineage, col-lineage, table-dependencies, acl-dependencies | | Orchestration | dag-generate, dag-deploy | | Validation | validate, freshness, test |

How to Navigate

1. Check {planning_artifacts}/ for existing planning documents.
2. Check {implementation_artifacts}/ for existing implementation work.
3. Recommend the next logical workflow based on what exists and what's missing.
4. If the user is unsure, suggest starting with domain discovery.

Outcome

Clear guidance on the appropriate next step in the Starflow workflow.