Whenever you're ready, walk me through how you'd approach this.
What core capabilities does your agent have? Walk through the end-to-end tool-calling flow: tool registration, parsing, invocation, and returning results.
Asked in the round 1 stage. Explain the design as a practical end-to-end system, including how an agent discovers tools, selects and validates a tool call, executes it safely, and incorporates the result into the next model step. Discuss interfaces, errors, observability, and safeguards.