📄️ Post Tool Call
In AI-driven applications, ensuring seamless interaction between users, tools, and the LLM is crucial. To address the challenge of managing LLM behavior based on tool outputs, we introduce a technique called Post-Tool Call Prompting. This technique enables developers to adjust the LLM's behavior after a tool executes and before the results are presented to the user, allowing for specific flow control based on the outcomes of the tools used.
📄️ Rerun
In CMND.ai, where you can create custom functions to interact with the language model, the rerun attribute plays a crucial role in determining the availability of these functions after their initial execution. The rerun attribute allows developers to control whether a function remains available for repeated use or is removed from the list of available functions after it has been invoked once, or even allowing the tool to rerun but only with different parameters.
📄️ Memory Object
Introduction
📄️ Prerequisites
The prerequisites attribute in CMND's function configuration is used to specify dependencies that must be fulfilled before a tool can be executed. By defining a list of prerequisite tools, developers can enforce an execution order, ensuring that certain tools are run in a specific sequence to maintain logic and consistency within the application/conversation.