Commands
Overview
Commands follow the pattern ags <service> <resource> <method>. Each command maps to one AGS API endpoint. No command has hidden side effects beyond the operation named in its path.
Describe the API surface
ags describe
This returns a machine-readable map of all available commands and their parameters. AI agents use this output to discover what the AGS CLI can do before selecting a command or workflow to run.
Discover the surface
The AGS CLI follows two usage shapes:
ags [FLAGS] <COMMAND> [OPTIONS]
ags [FLAGS] <SERVICE> <RESOURCE> <METHOD> [OPTIONS]
--help works at every level. Use it to explore what is available without running anything:
ags --help
ags iam --help
ags iam users --help
ags iam users search --help
ags describe narrows by service, resource, or method and returns JSON that an AI agent or script can consume directly:
ags describe # outputs all services
ags describe iam # outputs resources within a service
ags describe iam users # outputs methods within a resource
ags describe iam users search # outputs full parameter schema for a method
ags doctor checks your environment, configuration, and connectivity. ags completions generates a shell completion script for your shell.
Common options
Every ags command accepts these options:
| Option | Description |
|---|---|
--format json | Return output as JSON instead of human-readable text. Use this in scripts. |
--no-input | Disable interactive prompts. Required for CI and non-interactive environments. |
Telemetry
The AGS CLI collects anonymous usage data by default. This data helps the team improve the tool.
To opt out, set the DO_NOT_TRACK environment variable before running any command.
On macOS and Linux:
export DO_NOT_TRACK=1
On Windows (PowerShell):
$env:DO_NOT_TRACK = "1"