---
title: "List an agent's schemas"
method: GET
path: "/rest/api/v1/agents/{agent_id}/schemas"
tags: ["Agents"]
---

# List an agent's schemas

`GET /rest/api/v1/agents/{agent_id}/schemas`

Return [agent](https://developers.glean.com/agents/agents-api)'s input and output schemas. You can use these schemas to detect changes to an agent's input or output structure.

## Path parameters

- `agent_id` string, required — The ID of the agent.

## Query parameters

- `locale` string
- `timezoneOffset` integer

## Headers

- `X-Glean-ActAs` string, email
- `X-Glean-Auth-Type` string

## Response `200`

Success

- AgentSchemas — Defines the structure and properties of an agent.
  - `agent_id` string, required — The ID of the agent.
  - `name` string — The name of the agent.
  - `input_schema` object, required — The schema for the agent input. In JSON Schema format.
  - `output_schema` object, required — The schema for the agent output. In JSON Schema format.
  - `tools` ActionSummary[] — List of tools that the agent can invoke. Only included when include_tools query parameter is set to true.
    - `tool_id` string, required — The unique identifier of the action.
    - `display_name` string, required — The display name of the action.
    - `type` string — The type of tool - RETRIEVAL for read-only operations, ACTION for operations that modify data.
    - `auth_type` string — The authentication type required - OAUTH_USER, OAUTH_ADMIN, API_KEY, BASIC_AUTH, DWD (domain-wide delegation), or NONE.
    - `write_action_type` string — For write actions only - REDIRECT (client renders URL) or EXECUTION (external server call).
    - `is_setup_finished` boolean — Whether this action has been fully configured and validated.
    - `data_source` string — Indicates the kind of knowledge a tool would access or modify. Company knowledge: - Glean search, and any native tools that derive from it (e.g., expert search, code search) - Native federated tools to company data sources (e.g., outlook search) World knowledge: - Platform action like bravewebsearch, geminiwebsearch, etc Neutral knowledge: - Native tools that don't access or modify content via APIs (e.g., file analyst, think) - Platform read or write tools (creator has to determine their knowledge implications)

## Other responses

- `400` — Bad request
- `403` — Forbidden
- `404` — Not Found
- `422` — Validation Error
- `500` — Internal server error

---

[API](https://skmtc.net/gleanwork/apis/glean-client-api.md) · [All operations](https://skmtc.net/gleanwork/apis/glean-client-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/gleanwork/glean-client-api/versions/98678e54a6c5/schema)