This site is obsolete and should be used for reference only. The information in this documentation is not guaranteed to work for Bot Framework SDK versions past 4.9.1.

Manifest

The Skill manifest enables Skills to be self-describing in that they communicate the name and description of a Skill. Each action provides utterances that the caller can use to identify when an utterance should be passed across to a skill along with slots (parameters) that it can accept for slot-filling if required.

This manifest provides all of the metadata required for a calling Bot to know when to trigger invoking a skill and what actions it provides. The manifest is used by the Botskills command-line tool to configure a Bot to make use of a Skill.

Each skill exposes a manifest endpoint enabling easy retrieval of a manifest, this is typically found at the /manifest/manifest.json of your Skill URI.

Manifest structure

A manifest is made up of the following structure:

  • Description
  • Endpoints
  • DispatchModels
  • Activities
  • Definitions
  • ResponseObject

Description

The top level section of your Manifest provides high level information relating to your Skill, the table below provides more information on each item.

Parameter Description Required
id Identifier for your skill, no spaces or special characters Yes
name Display name for your skill Yes
description Description of the capabilities your Skill provides Yes
publisherName Publisher name Yes
version Version number for your skill Yes
iconUrl Icon Uri representing your skill, potentially used to show the skills registered with a Bot No
copyright Copyright message No
license License information No
privacyUrl Exposed entrypoint for communicating with your skill No
tags Exposed entrypoint for communicating with your skill No
{
  "$schema": "https://schemas.botframework.com/schemas/skills/v2.0/skill-manifest.json",
  "$id": "SampleSkill",
  "name": "SampleSkill",
  "description": "SampleSkill description",
  "publisherName": "Your Company",
  "version": "1.0",
  "iconUrl": "https://{YOUR_SKILL_URL}/SampleSkill.png",
  "copyright": "Copyright (c) Microsoft Corporation. All rights reserved.",
  "license": "",
  "privacyUrl": "https://{YOUR_SKILL_URL}/privacy.html",
  "tags": [
    "sample",
    "skill"
  ]
}

Endpoints

The endpoints section details 1 or more endpoints that your Skill will accept messages from.

  • endpointUrl must be manually updated to reflect the deployed location of your Skill.
  • msAppId must be manually updated to reflect the Azure AD Application ID of your deployed skill, this can be found in your appSettings.json file.
"endpoints": [
  {
    "name": "production",
    "protocol": "BotFrameworkV3",
    "description": "Production endpoint for the SampleSkill",
    "endpointUrl": "https://{YOUR_SKILL_URL}/api/messages",
    "msAppId": "{YOUR_SKILL_APPID}"
  }
]

Dispatch Models

The dispatchModels section provides pointers to accompanying language understanding data enabling a caller to train a local dispatcher to identify utterances that should be routed to the skill.

  • languages section supports multiple locales enabling a Skill to surface LU sources for each supported language.
  • url property enables the manifest to point at a variety of locations including file, http and luis endpoints.
  • intents section provides an optional mapping of Intents to the default message routing address. This enables a caller to only retrieve the LU data for the supported intents rather than the entire LU model.
"dispatchModels": {
  "languages": {
    "en-us": [
      {
        "id": "SampleSkillLuModel-en",
        "name": "SampleSkill LU (English)",
        "contentType": "application/lu",
        "url": "",
        "description": "English language model for the skill"
      }
    ]
  },
  "intents": {
    "SampleIntent": "#/activities/message",
  }
}

Activities

The activities section optionally defines a set of actions that the Skill supports. In assistant scenarios, utterance based triggering is typically used whereby an utterance is passed across to the Skill for intent detection and processing - the client’s responsibility is purely to detect the question is within the domain of a Skill.

Action based invocation is analagous to a function call. It enables the caller to invoke a specific capability of a Skill optionally passing input data (slots) in the form of an input object (value) and receiving data back through the form of an output object (resultValue). This invocation is performed through an Activity of type Event with the Name property set to the required Action name.

The Value property is used on the incoming event used to indicate an action and also the final EndOfConversation activity sent from the Skill to a calling Bot.

If data isn’t provided, the Skill can prompt for missing information as usual using Prompts. A Skill can also send response activities in addition to any result object.

The example activity definition below shows a SampleAction action being defined with value and resultValue types being specified in addition to a general message handler.

"activities": {
    "sampleAction": {
      "description": "Sample action which accepts an input object and returns an object back.",
      "type": "event",
      "name": "SampleAction",
      "value": {
        "$ref": "#/definitions/inputObject"
      },
      "resultValue": {
        "$ref": "#/definitions/responseObject"
      }
    },
    "message": {
      "type": "message",
      "description": "Receives the users utterance and attempts to resolve it using the skill's LU models"
    }
  }

Definitions

The definitions section provides the definitions for any types referenced in the preceding activities section. An example below defines two example objects.

"definitions": {
    "inputObject": {
      "type": "object",
      "properties": {
        "name": {
          "type": "string",
          "description": "The users name."
        }
      }
    },
     "responseObject": {
      "type": "object",
      "properties": {
        "customerId": {
          "type": "integer",
          "description": "A customer identifier."
        }
      }
    }
}

Example Skill Manifest

{
  "$schema": "https://schemas.botframework.com/schemas/skills/v2.0/skill-manifest.json",
  "$id": "SampleSkill",
  "name": "SampleSkill",
  "description": "Sample Skill description",
  "publisherName": "Your Company",
  "version": "1.0",
  "iconUrl": "https://{YOUR_SKILL_URL}/sampleSkill.png",
  "copyright": "Copyright (c) Microsoft Corporation. All rights reserved.",
  "license": "",
  "privacyUrl": "https://{YOUR_SKILL_URL}/privacy.html",
  "tags": [
    "sample",
    "skill"
  ],
  "endpoints": [
    {
      "name": "production",
      "protocol": "BotFrameworkV3",
      "description": "Production endpoint for the Sample Skill",
      "endpointUrl": "https://{YOUR_SKILL_URL}/api/messages",
      "msAppId": "{YOUR_SKILL_APPID}"
    }
  ],
  "activities": {
    "sampleAction": {
      "description": "Sample action which accepts an input object and returns an object back.",
      "type": "event",
      "name": "SampleAction",
      "value": {
        "$ref": "#/definitions/inputObject"
      },
      "resultValue": {
        "$ref": "#/definitions/responseObject"
      }
    },
    "message": {
      "type": "message",
      "description": "Receives the users utterance and attempts to resolve it using the skill's LU models"
    }
  },
  "definitions": {
    "inputObject": {
      "type": "object",
      "properties": {
        "name": {
          "type": "string",
          "description": "The users name."
        }
      }
    },
     "responseObject": {
      "type": "object",
      "properties": {
        "customerId": {
          "type": "integer",
          "description": "A customer identifier."
        }
      }
    }
  }
}