Custom packages and custom connectors in prompts - Prompt Co-Pilot for Automators to automatically detect and map custom packages and custom connectors. Accelerate development, enforce standardized logic, and reducing maintenance across Automation 360 workflows.

The administrator enables custom packages for Co-Pilot from the Control Room admin panel. A JSON file can be uploaded that details the allowed list of packages/connectors you want Co-Pilot to use.

The following is an example of the JSON schema.

{
  "custom_packages" : [ {
    "name" : "AAI Enterprise Knowledge",
    "description" : "This package enables comprehensive interaction with an Enterprise Knowledge Base, supporting resource management, querying, conversational AI, and content generation. It facilitates adding, removing, and retrieving resources, executing queries with or without follow-ups, engaging in document/image-based chats, creating projects, and generating images from text.",
    "usage" : "Use this package to manage and query an Enterprise Knowledge Base. It supports adding files/URLs, removing resources, retrieving statuses, creating projects, and interacting via text, documents, or images. Ideal for automating RAG workflows, content extraction, and AI-driven responses within enterprise knowledge systems.",
    "version" : "1.2.1-20250708-173520",
    "commands" : {
      "AddFilesToKnowledgeBase" : "Uploads local files to the Knowledge Base, associating them with a specified project and optional metadata. Returns a list indicating the status and any errors for each file.",
      "AddUrlsToKnowledgeBase" : "Adds online resources via URLs to the Knowledge Base, linked to a specified project with optional metadata. Returns a list showing the status and any errors for each URL.",
      "AskQuestion" : "Submits a user query to the Knowledge Base using a specified agent and project, optionally targeting specific documents. Returns a dictionary containing the JSON response.",
      "AskQuestionWithFollowUps" : "Sends a query to the Knowledge Base with the option to continue a previous conversation. Uses specified agent, project, and documents. Returns a dictionary with the JSON response.",
      "ChatWithDocument" : "Processes a local document to extract information based on a user query using a specified agent and project. Returns a dictionary containing the JSON response.",
      "ChatWithImage" : "Processes local image files to extract information based on a user query using a specified agent and project. Returns a dictionary containing the JSON response.",
      "CreateNewProject" : "Creates a new project in the Knowledge Base with specified name, description, type, and temperature settings. Returns the project ID of the newly created project.",
      "DeleteChat" : "Deletes a specified chat from the Knowledge Base within a given project. Returns a list indicating the status and any errors for each deletion.",
      "TextToImage" : "Generates an image from provided text using a specified agent and project, saving it to a local folder. Returns a dictionary containing the JSON response.",
      "GetAllResource" : "Retrieves all resources uploaded to the Knowledge Base for a specified project. Returns a list of resource identifiers.",
      "GetFileStatus" : "Checks the synchronization status of a specified file in the Knowledge Base for a given project. Returns a string indicating whether the file is synced or failed.",
      "RemoveFromKnowledgeBase" : "Removes specified files or URLs from the Knowledge Base within a given project. Returns a list indicating the status and any errors for each removal."
    },
    "conditionals" : { },
    "iterators" : { },
    "instructions" : [ ]
  }],
  "settings" : {
    "package_conflict_selection_priority" : "system",
    "no_match_policy" : {
      "Windows": "none",
      "Mac": "None"
    }
  }
}