Can I enable working memory but prevent the LLM from automatically updating it? I want the agent to read working memory but only update it programmatically via memory.updateWorkingMemory().
Right now, workingMemory.enabled: true automatically adds the update-working-memory tool. Is there a way to disable the tool while keeping working memory enabled? I am currently using a thread-scoped schema based working memory, but whenever the LLM is trying to update the working memory it fails quite often. I could update it way more reliably via code from my tools / workflows.