home / skills / sfc-gh-dflippo / snowflake-dbt-demo / devcontainer-setup

devcontainer-setup skill

needs review

This skill helps you create AI-ready DevContainers with snowflake-ai-tools scaffolding, enabling rapid containerized development across AI workflows.

npx playbooks add skill sfc-gh-dflippo/snowflake-dbt-demo --skill devcontainer-setup

Review the files below or copy the command above to add this skill to your agents.

Files (7)

SKILL.md

3.5 KB

---
name: devcontainer-setup
description:
  Create Universal DevContainers optimized for AI agentic workflows with Claude Code, Snowflake CLI,
  Cortex Code, and dbt. Use when setting up development containers, configuring devcontainer.json,
  scaffolding AI-ready environments, or when the user mentions devcontainers, containerized
  development, or Docker development environments.
---

# DevContainer Setup Skill

## When to Use

- User asks to set up a DevContainer or development environment
- User mentions AI-ready development setup
- Project needs containerized development with multiple language runtimes

## Instructions

### Step 1: Create directory structure

```bash
mkdir -p .devcontainer/snowflake-ai-tools
```

### Step 2: Copy templates

Copy these files from the skill templates folder:

| Source                                                   | Destination                                                  |
| -------------------------------------------------------- | ------------------------------------------------------------ |
| `templates/devcontainer.json`                            | `.devcontainer/devcontainer.json`                            |
| `templates/snowflake-ai-tools/devcontainer-feature.json` | `.devcontainer/snowflake-ai-tools/devcontainer-feature.json` |
| `templates/snowflake-ai-tools/install.sh`                | `.devcontainer/snowflake-ai-tools/install.sh`                |

If no `requirements.txt` exists in project root, also copy:

| Source                       | Destination        |
| ---------------------------- | ------------------ |
| `templates/requirements.txt` | `requirements.txt` |

### Step 3: Make scripts executable

```bash
chmod +x .devcontainer/snowflake-ai-tools/install.sh
```

### Step 4: Remind user to create host directories

Run this on their host machine before opening the devcontainer:

```bash
mkdir -p ~/.m2 ~/.npm ~/.cache/uv ~/.cargo ~/.snowflake ~/.claude ~/.cursor
```

### Step 5: Remind user about config files

These should exist on the host:

- `~/.snowflake/connections.toml` - Snowflake CLI config
- `~/.claude/` - Claude Code config

## Template Locations

All templates are in this skill's `templates/` folder:

- [templates/devcontainer.json](templates/devcontainer.json)
- [templates/snowflake-ai-tools/](templates/snowflake-ai-tools/)
- [templates/requirements.txt](templates/requirements.txt)

## Troubleshooting

| Issue                      | Solution                                                                                  |
| -------------------------- | ----------------------------------------------------------------------------------------- |
| NPM EACCES errors          | `sudo chown -R $(whoami) "$HOME/.npm"`                                                    |
| UID mismatch on Linux      | Verify `updateRemoteUserUID: true` in devcontainer.json                                   |
| Cache permission issues    | Run `snowflake-ai-tools-setup` to re-fix permissions                                      |
| Mount directory not found  | Run `mkdir -p ~/.m2 ~/.npm ~/.cache/uv ~/.cargo ~/.snowflake ~/.claude ~/.cursor` on host |
| Tools not found            | Check Docker build logs for install errors                                                |
| Different user than vscode | Update mount targets and containerEnv paths in devcontainer.json                          |

## Reference

See [README.md](README.md) for detailed documentation on:

- What's included (features, tools, mounts)
- Compatibility (base images, architectures)
- Python environment setup

Overview

This skill creates Universal DevContainers optimized for AI agent workflows, bundling Claude Code, Snowflake CLI, Cortex Code, and dbt. It scaffolds the .devcontainer layout, copies ready-made templates, and prepares host mounts and config hints so containers work smoothly for data and model development. Use it to get a consistent, reproducible developer environment for agentic and data engineering tasks.

How this skill works

The skill scaffolds a .devcontainer directory and installs a feature pack that includes Snowflake and AI tooling. It copies devcontainer.json, a feature manifest, and an install script into the project, makes the scripts executable, and instructs the user to create host-side cache and config directories for seamless mounts. Finally, it provides common troubleshooting checks for permission and mount issues.

When to use it

You need a reproducible containerized dev environment for AI agents or data engineering
Setting up devcontainer.json or adding features like Snowflake CLI, Claude, or dbt
Onboarding new contributors who require the same toolchain and mounts
Preparing development for multi-language projects (Python, Node, Java, Rust)
Resolving inconsistent local setups caused by missing host caches or config files

Best practices

Create the .devcontainer/snowflake-ai-tools directory and copy templates before opening the container
Run chmod +x on install scripts so the container build can execute them
Create host directories (~/.m2, ~/.npm, ~/.cache/uv, ~/.cargo, ~/.snowflake, ~/.claude, ~/.cursor) to avoid mount failures
Keep Snowflake CLI and Claude config files on the host (~/.snowflake/connections.toml and ~/.claude/) rather than in the container
Set updateRemoteUserUID: true in devcontainer.json on Linux hosts to avoid UID mismatches

Example use cases

Scaffold a devcontainer for a dbt + Snowflake analytics project that also needs Claude Code for local agent testing
Provision a development environment for a multi-runtime repo (Python, Node, Java) that shares caches across containers
Onboard a new team member with a single command to copy templates and create required host folders
Fix permission and tool-install issues by re-running the included snowflake-ai-tools setup script

FAQ

What host directories must exist before opening the devcontainer?

Create ~/.m2, ~/.npm, ~/.cache/uv, ~/.cargo, ~/.snowflake, ~/.claude, and ~/.cursor on your host to ensure mounts are available.

Where should Snowflake and Claude configs live?

Keep Snowflake CLI config at ~/.snowflake/connections.toml and Claude Code configuration in ~/.claude/ on the host for secure, persistent access.