agents-clilisted

# Agents CLI Use Google’s `agents-cli` to scaffold, run, evaluate, and deploy agents on Google Cloud; [google-adk](../google-adk/SKILL.md) owns Python ADK implementation inside that project. ## Workflow 1. **Install only the Python CLI** and inspect its current contract before scaffolding. Keep the compatible release pin aligned with the upstream skills you review. ```bash uv tool install "google-agents-cli" agents-cli --version agents-cli --help ``` 1. **Review the official skills** before using them. List the bundle and read each selected snapshot without installing it. ```bash skills add google/agents-cli --list skills use google/agents-cli@<name> ``` 1. **Default to the full ADK scaffold** with Agent Runtime, Gateway-ready image, BigQuery analytics, GitHub Actions, Terraform and `AGENTS.md`. The CLI creates the project directory; `--skip-checks --yes` avoids interactive cloud setup. Rehearse in a fresh temporary directory using [scaffold qualification](references/scaffold-qualification.md) when the CLI version or option set changes. ```bash agents-cli create <name> --agent adk \ --agent-guidance-filename AGENTS.md --agent-gateway --bq-analytics \ --deployment-target agent_runtime --cicd-runner github_actions \ --skip-checks --yes cd <name> skills add google/agents-cli --skill <name> -y skills list ``` Running `skills add` after `cd` keeps the reviewed skill project-scoped; use the immutable reviewed snapsho