Show command hierarchy in `agentstack --help`

[penge]

Problem

The Agent Stack CLI already supports a structured command hierarchy such as:

agentstack agent run
agentstack agent logs

However, the hierarchy is not visible in the top-level help output. Many agent-related commands appear directly at the top level, making the CLI appear flat.

Current agentstack --help output

Usage: agentstack [OPTIONS] COMMAND [ARGS]...

╭─ Getting Started ──────────────────────────────────────────────────────────╮
│ ui       Launch the web interface                                          │
│ list     View all available agents                                         │
│ info     Show agent details                                                │
│ run      Run an agent interactively                                        │
╰────────────────────────────────────────────────────────────────────────────╯

╭─ Agent Management [Admin only] ────────────────────────────────────────────╮
│ add                               Install an agent (Docker, GitHub)        │
│ remove                            Uninstall an agent                       │
│ update                            Update an agent                          │
│ logs                              Stream agent execution logs              │
│ env                               Manage agent environment variables       │
│ build                             Build an agent remotely                  │
│ client-side-build                 Build an agent container image locally   │
╰────────────────────────────────────────────────────────────────────────────╯

╭─ Platform & Configuration ─────────────────────────────────────────────────╮
│ connector       Manage connectors to external services                     │
│ model           Configure 15+ LLM providers [Admin only]                   │
│ platform        Start, stop, or delete local platform [Local only]         │
│ server          Connect to remote Agent Stack servers                      │
│ user            Manage users and roles [Admin only]                        │
│ self version    Show Agent Stack CLI and Platform version                  │
│ self upgrade    Upgrade Agent Stack CLI and Platform [Local only]          │
│ self uninstall  Uninstall Agent Stack CLI and Platform [Local only]        │
╰────────────────────────────────────────────────────────────────────────────╯

Because the agent command group is not visible in the help output, commands such as run, list, logs, and info appear to operate directly at the top level of the CLI.

As the CLI grows and additional command groups are introduced (model, connector, platform, server), commands like logs or list may become ambiguous.

Proposal

Expose the command hierarchy in the help output so the CLI structure is immediately visible.

Example:

Usage: agentstack [OPTIONS] COMMAND [ARGS]...

╭─ Core Commands ────────────────────────────────────────────────────────────╮
│ agent      Manage agents (run, install, inspect, logs)                     │
│ model      Manage model providers [Admin only]                             │
│ connector  Manage connectors to external services                          │
│ platform   Manage Agent Stack platform [Local only]                        │
│ server     Manage Agent Stack servers and authentication                   │
╰────────────────────────────────────────────────────────────────────────────╯

╭─ Utility Commands ─────────────────────────────────────────────────────────╮
│ ui         Launch the web interface                                        │
│ upgrade    Upgrade Agent Stack CLI and Platform [Local only]               │
│ uninstall  Uninstall Agent Stack CLI and Platform [Local only]             │
╰────────────────────────────────────────────────────────────────────────────╯

Short commands such as:

agentstack run
agentstack list
agentstack logs

could still remain available as aliases for convenience and onboarding.

Benefits

• Makes the CLI structure easier to discover
• Aligns help output with the actual command hierarchy
• Reduces ambiguity around commands like run, logs, and list
• Scales better as additional command groups are introduced