adk-infra-expert
Use when provisioning Vertex AI ADK infrastructure with Terraform. Trigger with phrases like "deploy ADK terraform", "agent engine infrastructure", "provision ADK agent", "vertex AI agent terraform", or "code execution sandbox terraform". Provisions Agent Engine runtime, 14-day code execution sandbox, Memory Bank, VPC Service Controls, IAM roles, and secure multi-agent infrastructure.
When & Why to Use This Skill
The Adk Infra Expert skill automates the provisioning of production-grade Vertex AI Agent Development Kit (ADK) infrastructure using Terraform. It streamlines the setup of secure multi-agent environments, including Agent Engine runtimes, code execution sandboxes, and VPC Service Controls, ensuring enterprise-level security and compliance on Google Cloud Platform (GCP).
Use Cases
- Enterprise Agent Deployment: Rapidly provision standardized Vertex AI environments with pre-configured Terraform modules that include VPC Service Controls for data exfiltration protection.
- Secure Sandbox Creation: Automatically set up 14-day isolated code execution sandboxes to safely test agentic workflows and custom code execution.
- IAM and Security Governance: Generate least-privilege IAM roles and service accounts specifically tailored for Agent Engine and Memory Bank components.
- Infrastructure Validation: Use the skill to validate existing Terraform configurations against Google Cloud best practices for AI agent infrastructure.
| name | adk-infra-expert |
|---|---|
| description | | |
| allowed-tools | Read, Write, Edit, Grep, Glob, Bash(terraform:*), Bash(gcloud:*) |
| version | 1.0.0 |
| author | Jeremy Longshore <jeremy@intentsolutions.io> |
| license | MIT |
Adk Infra Expert
Overview
Provision production-grade Vertex AI ADK infrastructure with Terraform: secure networking, least-privilege IAM, Agent Engine runtime, Code Execution sandbox defaults, and Memory Bank configuration. Use this skill to generate/validate Terraform modules and a deployment checklist that matches enterprise security constraints (including VPC Service Controls when required).
Prerequisites
Before using this skill, ensure:
- Google Cloud project with billing enabled
- Terraform 1.0+ installed
- gcloud CLI authenticated with appropriate permissions
- Vertex AI API enabled in target project
- VPC Service Controls access policy created (for enterprise)
- Understanding of Agent Engine architecture and requirements
Instructions
- Initialize Terraform: Set up backend for remote state storage
- Configure Variables: Define project_id, region, agent configuration
- Provision VPC: Create network infrastructure with Private Service Connect
- Set Up IAM: Create service accounts with least privilege roles
- Deploy Agent Engine: Configure runtime with code execution and memory bank
- Enable VPC-SC: Apply service perimeter for data exfiltration protection
- Configure Monitoring: Set up Cloud Monitoring dashboards and alerts
- Validate Deployment: Test agent endpoint and verify all components
Output
Error Handling
See {baseDir}/references/errors.md for comprehensive error handling.
Examples
See {baseDir}/references/examples.md for detailed examples.
Resources
- Agent Engine: https://cloud.google.com/vertex-ai/generative-ai/docs/agent-engine/overview
- VPC-SC: https://cloud.google.com/vpc-service-controls/docs
- Terraform Google Provider: https://registry.terraform.io/providers/hashicorp/google/latest
- ADK Terraform examples in {baseDir}/examples/