Automate prompt chaining on LLMs with multi-agent workflow | MindPal Blog