Self Healing AI Agent > Introduction to Self Healing AI Agent recipe > Self Healing AI Agent recipe contents
  

Self Healing AI Agent recipe contents

The Self Healing AI Agent recipe contains an embedding model connection, a model connection, tool connections, and agent flows.
The following table lists the assets that the Self Healing AI Agent recipe contains:
Asset Name
Asset Type
Description
CheckFileExistence
Tool connection
Tool that receives a file name as input and executes a script to check whether the file exists. If the file exists, the tool returns code 0. Otherwise, it returns code 1. This tool works only for Windows remote servers.
CheckSuccessfulRun
Tool connection
Tool that connects to the IDMC environment and checks whether any successful execution occurred for a failed job after the event failure.
ConnectionTest
Tool connection
Tool that connects to the IDMC environment and tests the connection object. It takes a failed taskflow name as input and retrieves the source and target connection IDs from the idmc_connections.json file used by the taskflow. The JSON file contains a list of taskflows along with their source and target connection IDs.
CreateINC
Tool connection
Tool that creates an incident in a ServiceNow instance. It receives an assignment group, short description, and description as inputs. It returns a detailed JSON response that includes status codes other than 200 or 201 to indicate failures.
EmailSendProcess
Tool connection
Tool that sends emails by taking inputs like email IDs, subject, and body content, and then triggers the Application Integration process to send an email. If the email is triggered successfully, it returns a 200 OK response code. You can pass multiple email IDs separated by commas.
FetchInc
Tool connection
Tool that receives an incident number as input, fetches the details from ServiceNow based on that number, and returns the response as output.
FetchTFRunID
Tool connection
Tool that fetches details of the failed taskflow within a specified time window. It takes the taskflow name, start time, and end time as inputs and returns a single run ID from the most recent failed execution. If the start time and end time are not provided, it defaults to the last 6 hours from the current time.
FileListTool
Tool connection
Tool that lists all the available files in a directory path provided as input.
FileRenameTool
Tool connection
Tool that receives a file name (not the complete path) as input and executes commands to rename the file. If the file is renamed successfully, the tool returns code 0. Otherwise, it returns code 1.
GetErrorMessage
Tool connection
Tool that retrieves the error message and downloads the taskflow log file to a specific directory on the server. It takes the taskflow name (assetName) and taskflow run ID (runId) as inputs from the prompt or the failure event.
GetLoginToken
Tool connection
Tool that receives an identifier (cred_file_id) as input, fetches credentials from the <identifier>_creds.py file on the system, and connects to the IDMC environment to retrieve the session ID, organization name, organization ID, and base server URL.
RestartTaskflow
Tool connection
Tool that executes a taskflow using the IDMC REST API. It takes the taskflow name as input.
SlackConnection
Tool connection
Tool that posts messages to a Slack channel.
SOP_Manual
Tool connection
Pinecone vector database retriever tool that reads the contents of the SOP document.
UpdateInc
Tool connection
Tool that updates an incident in ServiceNow. It can update work notes or mark an incident as closed or completed. To update work notes, it takes the incident sys_id, work notes, and the action UPDATE as inputs. To close an incident, it takes the sys_id, resolution notes, and the action CLOSE as inputs.
UpsertVectorDB
Tool connection
Tool that takes text input, creates embeddings using the Azure OpenAI Embedding model, and upserts them into a vector database using Pinecone as the target.
IPS_AI_EmbedModel
Embedding model
OpenAI embedding model that creates embeddings from text.
Azure_OAI_IPS_SL
Model connection
Connects to the Azure OpenAI GPT4.1 model.
MonitoringAgent
Agent flow
Self-healing agent that analyzes the error messages, reads SOP documents, performs prerequisite and remediation actions for documented failures, and escalates by creating incidents for undocumented errors.
SOP Governance Agent
Agent flow
Agent that analyzes resolved incidents and updates the SOP with new content based on approvals.