Bug Description

Rerunning history on workflows with sensitive data, the values show up as x_secret_name in the browser playbacks.

Reproduction Steps

1. Run an agent with sensitive data and get history = await agent.run(), and see the correct values placed
2. Use agent.rerun_history(history)
3. See the Agent outputs now include ... tags

Code Sample

agent = Agent(
            task="search for x_secret in google in a new tab",
            llm=model,
            sensitive_data=secret_vars,
        )

        history = await agent.run()
        result = history.final_result()
        rerun_result = await agent.rerun_history(history, delay_between_actions=0.1)

Version

0.1.36

LLM Model

GPT-4o

Operating System

macos 15.4

Relevant Log Output