To date there has been no try to supply a comprehensive taxonomy of those agent applications. This kind of taxonomy could enable actors across the AI source chain to a lot more Obviously share details about system abilities and issues. For example, This tends to empower an AI agent developer to share tool abilities and restrictions with downstream