list_prompts

List every customer prompt this credential can see. Depending on its context, that can include the represented user's own prompts, team-owned prompts from its bound teams, and prompts it can reach in the organization's shared catalogue. Use it to discover what already exists before you read, edit, or copy a prompt. It takes no parameters.

What you can see is broader than what you can change: a prompt you can reach in the shared catalogue shows up here so you can read and copy it, even if only a person can edit the catalogue itself.

Parameters

None.

Response

Field	Type	Description
prompts	array	The prompts you can see (see below).

Each entry:

Field	Type	Description
name	string	Exact caller-relative address to pass to get_prompt.
description	string	The prompt's description, when it has one.

In an organization, the represented member's private prompts are listed with a me/ prefix, and prompts owned by a bound team use the team's name (marketing/launch-copy), so you can tell them apart from the shared catalogue. A userless organization credential has no me/ entries. The shared catalogue stays unprefixed. Pass the name exactly as shown to get_prompt; only private (me/) and team addresses are writable in organization scope.

Example

Result:

{
	"prompts": [
		{ "name": "launch-copy", "description": "Product launch announcement" },
		{ "name": "marketing/tone-guide", "description": "Team voice and tone" },
		{ "name": "support-triage" }
	]
}

Errors

Each failure carries a precise code in the response. A capability denial is returned as a tool result with isError: true, so the agent can read the code and self-correct; a server fault is a JSON-RPC protocol error.

Code	Cause	Delivered as
capability_denied	The credential lacks the Prompts capability required for this read.	tool error (isError)
internal_error	An unexpected error occurred.	protocol error (internal_error)

Notes

• The list mirrors the customer-owned part of native prompts/list. Platform prompts such as apistash/agent-doctor are available only through the native prompt methods and are excluded from management tools.
• To read a prompt's full text, pass its name to get_prompt.
• Listing counts as one tool call against your plan's tool-invocation allowance.