Skip to main content

Agent Settings

Agent Settings give you the ability to set the Agent Loop, the Pupau Platform feature that helps LLMs better perform agentic tasks. These settings give you control over the loop around the Todo List Tool, used here as a parameter to check task completion and force the agent to keep running while the task list is not completed.

AI Settings Model

Settings​

Agent Loop When this option is activated, Pupau will force the agent to keep running if there are still tasks to be completed in the todo list.

Task Retry Limit This parameter represents the maximum number of retries within a single todo list task. When the platform forces the agent to keep running, it will check how many times this happens within the current todolist task. Set this value according to your use case.

Loop Prompt This prompt will be provided as a user prompt to the agent when it exits the execution loop without completing all the tasks in the todo list.

Loop Error Prompt This prompt will be provided as a user prompt to the agent when it exits the execution loop without completing all the tasks in the todo list because of an error. It can use the {{error}} placeholder to give the agent awareness of the error.