satsuma-lang

Write a new entry in "The Satsuma Diaries" — a funny, tongue-in-cheek project changelog written by Pip, a junior intern AI with strong SE London energy. Run this after a coding session or on a schedule. The skill decides whether the day merits its own entry or should be rolled into a quiet-days summary. Files entries under satsuma-diaries/yyyy/mm/yyyy-mm-dd.md and tracks progress in satsuma-diaries/.progress.json.

Generate OpenLineage events from Satsuma (.stm) mapping specs. Use this skill whenever the user wants to export Satsuma lineage to OpenLineage, integrate Satsuma with Marquez or DataHub, generate OpenLineage JSON from .stm files, create lineage events for a data catalog, or bridge Satsuma specs into an OpenLineage-compatible lineage backend. Also trigger for requests like "export lineage to OpenLineage", "generate Marquez events", "create column-level lineage JSON", "integrate Satsuma with DataHub/Atlan/OpenMetadata", or "publish lineage from my mapping spec". Requires the `satsuma` CLI for structural extraction. Generates spec-compliant OpenLineage 2-0-2 RunEvents with standard and custom facets.

Convert Excel mapping spreadsheets into idiomatic Satsuma .stm files. Use when the user has an .xlsx file containing source-to-target data mapping definitions and wants to generate Satsuma output. Three-phase workflow: survey the spreadsheet structure, generate Satsuma with chunked extraction, then critique and refine.

Explain Satsuma (.stm) data mapping files in plain English for business stakeholders, analysts, and non-technical reviewers. Use this skill whenever the user asks to explain, summarize, review, walk through, or describe a Satsuma file or workspace — including requests like "what does this mapping do", "explain this to my PM", "review this spec", "what are the risks", "is anything missing", "onboarding summary", or "walk me through this". Also trigger when users ask for a data quality assessment, PII audit summary, coverage check, or impact analysis of a .stm file. This skill requires the `satsuma` CLI to be installed and available on PATH.

Reverse-engineer Satsuma (.stm) mapping specs from an existing dbt project. Use this skill whenever the user wants to generate Satsuma from dbt, convert a dbt project to Satsuma, extract mapping specs from dbt models, document dbt transformations in Satsuma format, or create a Satsuma workspace from their existing dbt codebase. Also trigger when the user says things like "I have a dbt project and want to try Satsuma", "convert my dbt models to .stm", or "reverse-engineer my dbt lineage into Satsuma". Requires access to the dbt project files on disk. The `satsuma` CLI should be installed for validation and formatting of the output.

Generate realistic synthetic test data from Satsuma (.stm) schema definitions. Use this skill whenever the user wants to create test data, sample data, fake data, seed data, mock data, or synthetic data from a Satsuma spec. Also trigger for requests like "generate CSV from this schema", "create test fixtures for my mapping", "I need sample data to test my pipeline", "mock up some data for this spec", or "seed data for my warehouse". Generates data that respects all schema constraints (types, enums, required, PII patterns, formats, defaults, filters) and maintains referential integrity across schemas connected by mappings. Supports CSV, JSON/JSONL, SQL INSERT, and Parquet output. For exotic source formats (COBOL copybooks, HL7, EDI, ISO 8583, etc.), this skill generates the logical data and suggests the user create a format-specific skill for physical serialisation. Requires the `satsuma` CLI.

Generate idiomatic dbt project scaffolds from Satsuma (.stm) mapping specs. Use this skill whenever the user wants to create a dbt project from Satsuma, generate dbt models from .stm files, scaffold a dbt pipeline from a mapping spec, or convert Satsuma to dbt. Also trigger for requests like "generate dbt from this spec", "create staging and mart models from this mapping", "scaffold a dbt project for this Satsuma file", or "turn my .stm into dbt SQL". Handles Kimball star schemas, Data Vault 2.0, merge strategies, governance metadata, and consumer schemas (reports/models → dbt exposures). Requires the `satsuma` CLI for structural extraction. Generates dbt projects that use {{ ref() }} and {{ source() }} throughout — never hardcoded table names.

Convert Satsuma .stm files into professional, stakeholder-ready Excel workbooks. Deterministic, parser-backed export using the satsuma CLI for structural extraction and openpyxl for workbook generation. Produces Overview, Issues, Mapping, Schema, and Lookup tabs with full styling.

Review recent changes and determine whether a new Architectural Decision Record is warranted. If so, draft it (and mark any superseded ADRs) for user review before writing to adrs/.