examples: resolve invalid tool usage status code 400 if llm makes a mistake gpt-oss

[MarkWard0110]

The bug this fixes.
Sometimes the LLM can make a mistake in the tool's name. The client would output Tool x not found and submit a new chat request. The problem is that the chat request will contain the assistant tool request, but will be missing a tool response. This results in an HTTP status code 400, invalid tool usage.

The patch adds a tool response when the tool is not found. This will enable the client to inform the LLM that it performed a bad tool call as a tool response and avoid the HTTP 400 error.

Test with a prompt that will invoke many tool calls like I need the weather conditions for each state capital in the United States.

[ParthSareen]

Thanks!