aiml
SearchMy learning
Menu

Building, Evaluating, and Deploying Agents with Google Agents CLI

Google Agents CLI 1.11

Scaffold, enhance, upgrade, evaluate, deploy, publish and observe ADK agents across the full Google Agents CLI lifecycle.

Prerequisites
Python 3.12, a terminal, and basic familiarity with Google ADK agents (the Building Agents with Google ADK 2.0 course covers them).
Duration
about 2 hours of reading
Cost
Free, no sign-up
Start course

What you'll learn

  • Scaffold an ADK prototype with agents-cli scaffold create, keep App(name='app') aligned with agent_directory, pass agents-cli lint, and run multi-turn local sessions.
  • Graduate a prototype in place with scaffold enhance, run 3-way template merges with scaffold upgrade, and add or override CLI commands with agents-cli-extension.yaml.
  • Decouple trace collection (eval generate) from scoring (eval grade), synthesize multi-turn user simulations, cluster losses with eval analyze, and inspect result differences with eval compare and a separate regression gate.
  • Provision Terraform infrastructure, validate Agent Runtime and Cloud Run rollouts with deploy --dry-run, register ADK or A2A agents with publish gemini, and query two-tier BigQuery telemetry.

Use these courses for background and deeper implementation details.

Lessons

Module 1 · 2 lessonsScaffold, run, and extend agent projects

Create a prototype ADK project, run multi-turn sessions and A2A discovery locally, enhance with Terraform and CI/CD, and add project CLI extensions.

  1. Scaffold, lint, and run an ADK project with Google Agents CLIRead

    Scaffold a prototype ADK agent with agents-cli, inspect agents-cli-manifest.yaml, run agents-cli lint, and serve multi-turn ADK and A2A sessions locally.

    18 min readLabQuiz

  2. Enhance, upgrade, and extend an Agents CLI projectRead

    Add Cloud Run and CI/CD scaffolding in place with scaffold enhance, run 3-way template merges with scaffold upgrade, and wire agents-cli-extension.yaml.

    20 min readLabQuiz

Module 2 · 2 lessonsEvaluate, simulate, and compare agent runs

Generate traces from single-turn and continued (N+1) cases, score with local custom metrics and distinguish model-based grading, synthesize user simulations, and compare baseline vs candidate runs.

  1. Generate traces and grade evaluations with agents-cli evalRead

    Author single-turn and continued-conversation datasets, generate ADK traces with agents-cli eval generate, and score custom metrics in eval grade.

    22 min readLabQuiz

  2. Synthesize user simulations and compare evaluation runsRead

    Generate multi-turn user-simulated traces with agents-cli eval dataset synthesize, cluster losses with eval analyze, and diff runs with eval compare.

    20 min readLabQuiz

Module 3 · 2 lessonsDeploy, publish, and observe on Google Cloud

Coordinate Terraform with agents-cli deploy, register ADK or A2A agents in Gemini Enterprise, and configure two-tier Cloud Trace, GCS, and BigQuery completion telemetry.

  1. Provision Terraform infrastructure and deploy with agents-cliRead

    Separate Terraform infrastructure from code rollouts, validate Agent Runtime and Cloud Run rollouts with agents-cli deploy --dry-run, and inspect metadata.

    22 min readLabQuiz

  2. Publish to Gemini and configure agent observabilityRead

    Register a deployed agent with Gemini Enterprise in ADK or A2A mode using agents-cli publish gemini, and query two-tier GCS and BigQuery telemetry.

    20 min readLabQuiz