Developer Onboarding¶

Two options: devcontainer (recommended) or native setup.

Option A: Devcontainer (recommended)¶

Same container image as the AI agents. Zero drift, zero local install.

Prerequisites¶

Tool	Install
Docker / Podman	`brew install docker` or Docker Desktop
VS Code	With Dev Containers extension
GitHub CLI	`brew install gh`

Setup¶

# Clone a repo
gh repo clone dashecorp/conductor-e
cd conductor-e

# Open in VS Code → "Reopen in Container"
# Or from CLI:
devcontainer up --workspace-folder .

The devcontainer auto-configures: - Claude Code CLI + Codex CLI - Conductor-E hooks (automatic event reporting) - Git hooks (branch creation triggers WORK_STARTED) - gh CLI, git, curl, jq - Stack-specific tools (.NET, Python, or Node.js depending on repo)

Each repo has a .devcontainer/devcontainer.json pointing to the right stack image.

Available stack images¶

Image	Tools	Used by
`rig-agent-runtime:base`	Node.js 22, Claude CLI, gh CLI	rig-gitops, infra
`rig-agent-runtime:node`	+ TypeScript, Jest, ESLint	rig-agent-runtime
`rig-agent-runtime:dotnet`	+ .NET 10 SDK	conductor-e, dev-e, review-e
`rig-agent-runtime:python`	+ Python 3, pytest, black, ruff	Python projects

Option B: Native Setup¶

For macOS-native work (iOS development) or when you prefer not to use containers.

Prerequisites¶

Tool	Install
Claude Code	`npm install -g @anthropic-ai/claude-code`
GitHub CLI	`brew install gh`
gcloud CLI	`brew install google-cloud-sdk`
Node.js 22+	`brew install node`

1. Create Workspace¶

mkdir ~/repos/my-workspace && cd ~/repos/my-workspace

2. Clone Repos¶

# Rig infrastructure
for repo in conductor-e rig-agent-runtime rig-gitops dev-e review-e infra; do
  gh repo clone dashecorp/$repo repos/$repo
done

# iOS apps
for repo in nutri-e cutie star-rewards fast-e count-e drink-e heart-e; do
  gh repo clone Stig-Johnny/$repo repos/$repo
done

3. Install Rig Tools¶

Connects your development environment to the Conductor-E event pipeline (same as AI agents):

gh repo clone dashecorp/rig-tools repos/rig-tools
cd repos/rig-tools && ./install.sh && cd ../..

This installs conductor-e-hook CLI, configures Claude Code hooks, and sets up env vars. Works with any AI tool — see rig-tools README for details.

4. Set Up Claude Code¶

Copy the templates from rig-gitops/templates/:

cp repos/rig-gitops/templates/workspace-claude.md CLAUDE.md
cp repos/rig-gitops/templates/mcp.json.template .mcp.json

Edit CLAUDE.md: - Replace {NAME} with your workspace name - Replace {ROLE} with your role

Edit .mcp.json: - Fill in credential placeholders - Add additional MCP servers as needed (see templates/mcp.json.template for core set)

Migration note (2026-04-17): The memory MCP server key was renamed from rig-memory to memory in templates/mcp.json.template. If your local .mcp.json still has a rig-memory entry, rename the key to memory and update the config to use npx:
"memory": {
  "command": "npx",
  "args": ["-y", "@dashecorp/rig-memory-mcp"],
  "env": { "DB_URL": "…", "AGENT_ROLE": "dev-e" }
}
Earlier drafts of this note pointed at …/rig-memory-mcp/dist/index.js, which does not exist in the v2 package layout (plain ESM, index.js at root). See rig-gitops#67 for the background.

5. Cluster Access¶

# Authenticate to GCP
gcloud auth login
gcloud config set project invotek-github-infra

# Access the k3s cluster
gcloud compute ssh invotek-k3s --zone europe-north1-b
export KUBECONFIG=/etc/rancher/k3s/k3s.yaml

6. Key Resources¶

Resource	URL
Flux Dashboard	https://flux.dashecorp.com
Conductor-E API	https://conductor-e.dashecorp.com
Shared Workflow	`rig-gitops/AGENTS.md`
Doc Standard	`rig-gitops/docs/documentation-standard.md`

Workflow¶

Pick an issue from a repo's GitHub issues
Read the repo's AGENTS.md for build/test commands
Create feature branch: feature/issue-{N}-description
Implement with conventional commits
Update docs if behavior changed (YAML frontmatter required)
Create PR referencing Closes #{N}
Wait for CI + Review-E review
Squash merge to main