Orchestrate Extract→Transform→Value-map→Load→Represent data pipelines with CLI-driven skills.
npx skills add https://github.com/Soria-Inc/soria-stack --skill soria-stack
Please help me install this Agent Skill. Skill: soria-stack Source: https://github.com/Soria-Inc/soria-stack/tree/main Command: npx skills add https://github.com/Soria-Inc/soria-stack --skill soria-stack
Coordinating end-to-end data pipeline work is error-prone: teams accidentally write to prod, skip verification, or build dashboards without auditability. SoriaStack provides a consistent, CLI-first set of skills that enforce gates, provenance, and verification so analysts and engineers can move from scraping to production with confidence.
Start a session by ensuring the soria CLI and an active dev environment are available, run /env then /tools, and invoke the appropriate skill (for example, /ingest to build a pipeline or /dive to create a dive).
High-intent search queries and answers about installing and using this skill.
ETVLR data pipeline orchestration is managed through CLI-driven skills that execute Extract, Transform, Value-map, Load, and Represent stages with gated checkpoints and artifact outputs. You invoke specific commands like /ingest to build pipelines and /verify to run checks.
Environment isolation uses git worktrees paired with Neon branches and MotherDuck clones to create sandboxed dev and preview spaces. This mechanism prevents accidental prod writes by ensuring pipeline development and testing occur entirely within isolated database clones.
A headless browser is optional and only required for UI QA commands like /dashboard-review and /browse. Core pipeline tasks such as scraping, dbt mart creation, and verification checks run without it, relying instead on the soria CLI and active environment worktrees.
Preventing accidental production writes is enforced through artifact-based gates and mandatory prod write confirmations. The workflow requires promoting changes from isolated dev environments using git-based promotion, ensuring all transformations are verified before touching production.
Yes, you can diagnose and repair data quality issues by running the /diagnose command to identify problems and /ticket to track repairs. The pipeline applies verification checks via /verify to validate data marts before promotion.
ETVLR extends standard ETL by breaking downstream processing into Value-map, Load, and Represent stages, providing explicit checkpoints for schema mapping, dbt marts, and dashboard QA. You need it when coordinating end-to-end pipelines requires strict provenance, auditability, and verification gates.
Soria-IncManage software project environments via unified CLI for git, Neon, and MotherDuck.
Soria-IncRender dive manifests into markdown pivot tables via soria warehouse queries.
sigmacomputingMigrate Amazon QuickSight dashboards and datasets into Sigma Computing workbooks.
erichowensAutomate ETL/ELT pipeline design, validation, and operation across medallion architectures.
