When implementing tool orchestration for an agent that needs to dynamically select from multiple tools...

The decisive point is failure isolation: Option B keeps the agent’s decision path observable instead of burying behavior inside one prompt or one service. The stack-level anchor is clear: the Agent Toolkit model is to expose tools as reusable workflow components; that is what makes multi-tool agents testable under schema changes. The selected option specifically B states “LLM-based tool selection with structured tool descriptions and usage examples”, which matches the operational requirement rather than a superficial wording match. LLM-based selection works when tools have structured descriptions and schemas. Pure rules break when inputs are novel; randomness is indefensible in production. The runtime should therefore be built around schema-bound tool invocation, typed parameters, timeout envelopes, retry policy, and traceable function execution. The distractors fail because embedding tools inside the agent loop makes security review, timeout handling, and version control unnecessarily difficult. The answer is therefore about engineered control planes, not simply model capability. Schema validation, typed return objects, and trace IDs also make post-incident debugging realistic when a third-party dependency changes behavior.