What problem does it solve? Complex tasks like logic puzzles, math problems, and trade-off analysis often produce shallow or incorrect answers from LLM agents that respond immediately without thinking. This Skill teaches how to add explicit reasoning steps to Agno agents so they think before answering, improving answer quality on demanding tasks. ## Core Features & Use Cases - Reasoning models: Use natively reasoning-capable models (e.g., OpenAI o-series like o3-mini) that think internally before responding. - Managed chain-of-thought: Set reasoning=True on any Agno agent to let the framework orchestrate a reasoning cycle over any model, without switching providers. - ReasoningTools: Give the agent explicit think and analyze tools for transparent, inspectable reasoning steps, viewable with show_full_reasoning=True. - Use Case: Ask an agent to solve a word problem (e.g., two trains departing at different speeds) and compare the answer quality with and without reasoning enabled to decide which approach fits your workload. ## Quick Start Create an Agno agent with reasoning=True and ask it to solve a step-by-step logic problem while showing its full reasoning.
