Hey, I can help with this 👍 The issue is there’s no single “standard” doc for agent best practices yet... most people piece it together from patterns (prompt structure, tool usage, validation loops, etc.), so your agent doesn’t have a solid reference to follow. I’ve worked on structuring subagent prompts, when to use extract vs function nodes, and setting up clean validation flows ... happy to guide you through it. Quick question: what framework or stack are you building this on (LangChain, OpenAI Agents, custom, etc.)?