test-user-connector-action

Run one action of a user-defined connector you may edit, with real inputs, while you build it. Needs the fabric:skill.write grant.

curl -sS -X POST https://fabric.wexa.ai/mcp/$FABRIC_PROJECT_ID \
  -H "authorization: Bearer $FABRIC_API_KEY" \
  -H "Mcp-Session-Id: $MCP_SESSION_ID" \
  -H "content-type: application/json" -H "accept: application/json" \
  -d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"test-user-connector-action","arguments":{"connector_id":"6b1a3d5e7f9c0b2a4d6e8f02","action_key":"get_employee","inputs":{"employee_id":"E1"}}}}'

{
  "content": [
    {
      "type": "text",
      "text": "{\"lifecycle_id\":\"qlc_000017\",\"result\":{\"success\":true,\"output\":{\"id\":\"E1\",\"name\":\"Asha Rao\"}}}"
    }
  ],
  "isError": false
}

curl -sS https://fabric.wexa.ai/v1/test-user-connector-action \
  -H "authorization: Bearer $FABRIC_API_KEY" \
  -H "content-type: application/json" \
  -d '{"connector_id":"6b1a3d5e7f9c0b2a4d6e8f02","action_key":"get_employee","inputs":{"employee_id":"E1"}}'

{
  "lifecycle_id": "qlc_000017",
  "result": {
    "success": true,
    "output": {
      "id": "E1",
      "name": "Asha Rao"
    }
  }
}

import { Fabric } from '@wexa-fabric/sdk'

const fabric = new Fabric()
const result = await fabric.testUserConnectorAction({
  connector_id: '6b1a3d5e7f9c0b2a4d6e8f02',
  action_key: 'get_employee',
  inputs: {
    employee_id: 'E1'
  }
})

// result is the payload itself, with no `.result` to read
{
  success: true,
  output: {
    id: 'E1',
    name: 'Asha Rao'
  }
}

from wexa import Fabric

fabric = Fabric()
out = fabric.test_user_connector_action(
    connector_id="6b1a3d5e7f9c0b2a4d6e8f02",
    action_key="get_employee",
    inputs={
        "employee_id": "E1"
    },
)

# out is the result, already unwrapped
{
    "success": True,
    "output": {
        "id": "E1",
        "name": "Asha Rao"
    }
}

Parameters

Parameter	Type	Required	Notes
`action_key`	string	Required	The action, by its `key`. Must not be empty.
`connector_id`	string	Required	The user-defined connector, from `list-user-connectors`. An unknown id is refused (404). A missing id, or a connector you may not edit, is refused (502); the message names which.
`inputs`	object	Optional	The action's inputs, by name. Defaults to none.

Show full response

An example response, for an action that read one employee.

{
  "content": [
    {
      "type": "text",
      "text": "{\"lifecycle_id\":\"qlc_000017\",\"result\":{\"success\":true,\"output\":{\"id\":\"E1\",\"name\":\"Asha Rao\"}}}"
    }
  ],
  "isError": false
}

{
  "lifecycle_id": "qlc_000017",
  "result": {
    "success": true,
    "output": {
      "id": "E1",
      "name": "Asha Rao"
    }
  }
}

// the object you get back IS the result — there is no `.result` to read
{
  "success": true,
  "output": {
    "id": "E1",
    "name": "Asha Rao"
  }
}

// The TS SDK names these camelCase — lifecycleId, traceId — where Python uses
// lifecycle_id and trace_id. Both are non-enumerable: readable, but omitted by
// Object.keys() and JSON.stringify(). `result.lifecycleId` here is undefined.
result.lifecycleId // 'qlc_000017'

# a dict subclass; `result` is already unwrapped
{
 "success": True,
 "output": {
  "id": "E1",
  "name": "Asha Rao"
 }
}

out.lifecycle_id  # 'qlc_000017' — an attribute, not a key

Keeping the lifecycle_id is what lets you ask later why a call was allowed, refused or held.

More details

Response fields

success: Whether the action ran.
output: What the action returned, when success is true.
fault: When success is false: the status and detail the action failed with.

A live connector calls its API. An action that is not a GET really changes data there, so test it against data you can afford to change. A simulated connector returns simulated output.

A failed action is an answer, not an error: the call succeeds, with success set to false and the fault the action failed with.

The connector id in the example is illustrative. Swap it for one list-user-connectors returned to you.

Not marked consequential: it is the author's own test run of a connector they are building, and the default write gate does not hold it. It can still change data in the connector's API, so mind which action you run.

Every tool call works the same way: How every tool call works.