prepare-deployment

Configure profile-scoped credentials and production destinations for dltHub pipeline deployment.

Updated Jun 15, 2026
One-click install
npx skills add https://github.com/aminojagh/LLMZC --skill prepare-deployment-aminojagh
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: prepare-deployment
Source: https://github.com/aminojagh/LLMZC/tree/main/05_02_dlt_workshop/.claude/skills/prepare-deployment
Command: npx skills add https://github.com/aminojagh/LLMZC --skill prepare-deployment-aminojagh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires dlt, and includes scripts (resource) and references (resource) components.

What problem does it solve? Moving dlt pipelines from local development to the dltHub Platform requires separating dev and production credentials, configuring a production destination, and declaring deployable jobs — a process that is error-prone when done manually. ## Core Features & Use Cases - Profile-Scoped Secrets: Splits workspace secrets into dev and prod profile TOML files so each environment uses the right credentials. - Production Destination Setup: Configures named destinations (e.g., Motherduck instead of ephemeral duckdb) per profile, with a connectivity check script to verify access. - Deployment Manifest Creation: Builds __deployment__.py with @run.pipeline decorated functions and __all__ exports so the Runtime can discover jobs. - Use Case: You have a working dlt pipeline using local duckdb and want to deploy it to dltHub Runtime with Motherduck as the production warehouse while keeping local development unchanged. ## Quick Start Ask the assistant to prepare this dlt workspace for production deployment by splitting dev and prod secrets and configuring a Motherduck destination.

Frequently Asked Questions about prepare-deployment

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

FAQPage Schema
How do I deploy a dlt pipeline to production with dltHub?▼

Create profile-scoped secrets files like prod.secrets.toml, configure a production destination, and define a __deployment__.py manifest importing your decorated pipeline functions. Verify with dlthub deploy --dry-run before deploying.

How do I separate dev and prod credentials in dlt?▼

dlt supports profile-scoped TOML files such as dev.secrets.toml and prod.secrets.toml inside the .dlt directory. Profile files override workspace-level secrets.toml values when the corresponding profile is active.

Why does my duckdb data disappear after deploying to dltHub Runtime?▼

The dltHub Runtime erases ephemeral storage after runs, so local duckdb files do not persist. Use a persistent production destination like Motherduck, which keeps the duckdb-compatible workflow while storing data remotely.

What is a named destination in dlt?▼

A named destination is an alias that can resolve to different destination types per profile, such as duckdb on dev and Motherduck on prod. Pipelines reference the name, letting the same code run against different backends.

When do I need a __deployment__.py file in dltHub?▼

It is mandatory for workspaces with transformations, multiple pipelines, scheduled jobs, or followup triggers. Simple single-pipeline workspaces can use dlthub run directly without a deployment manifest.

How do I verify production destination credentials before deploying?▼

Run the check_destination.py script with the profile and destination name. It switches to the profile, opens a destination client connection, and reports whether the dataset exists, exiting nonzero on failure.