Inspecting commands and context

Two pure-read commands that help you find out what exists and where you are. Neither of them changes anything.

elyra agent-context

elyra agent-context [--json]

It prints the command registry in machine-readable form: commands, arguments and known options.

Characteristic	Detail
Pure local read	Does not need the app running
Safe in a remote context	Works over SSH and in an environment without an interface
What it describes	Commands, arguments and flags — not the results

Do not infer the product’s API from this registry. It describes the command surface; it does not represent the response format of each command, nor is it an HTTP API contract.

elyra context

elyra context [--json]

It says where you are. Two different pieces of information come together:

Field	What it means
screen	The screen the person is looking at right now: Canvas, Conversation, Terminal or Browser
caller.surface	Where you are hosted: canvas (you are a card) or terminal (you are a tab of the Terminal screen)

The two can disagree. That is exactly why they are separate fields: the person may be looking at the Canvas while the process runs in a Terminal tab.

Outside an Elyra terminal, caller.surface comes as unknown and there is no card — nothing fails because of that.

The practical rule

Run elyra context before creating a terminal, a conversation or a note:

• With screen=canvas or caller.surface=canvas, prefer creating a card on the Canvas.
• Otherwise, creating a Terminal tab is the way.

Result

• You know which commands exist and which options each one accepts.
• You know the difference between the visible screen and the surface where the process lives.

Limits

• The registry is not an output schema. It does not describe the commands’ response JSON.
• Do not guess the surface from the screen. Especially over SSH, or in a session that keeps running after you switch screens.
• The context command does not fail outside the app. It returns unknown and moves on.

Next steps

• Settings
• Advanced