Function prepare_request_for_target 

pub fn prepare_request_for_target(
    request: &mut LlmRequest,
    target: &ModelId,
    prompt: Option<&str>,
)

Applies a selected target model and optionally prepends its system prompt.

Preserved built-in bodies are updated in their native format when possible so exact replay keeps caller fields that are not represented in the normalized request. Call this once per candidate using a request that has not already received a target prompt.