fix: reset executor state between sequential task executions

[Anandesh-Sharma]

Summary

• Resets self.messages and self.iterations at the start of invoke() and ainvoke() in CrewAgentExecutor
• When one agent handles multiple sequential tasks, the executor is reused, causing messages from Task 1 to leak into Task 2's LLM context
• This also prevents premature max iterations limits from carrying over
• Matches the behavior already implemented in the experimental AgentExecutor

Fixes #4389

Test plan

• Create a Crew with one agent assigned to two sequential tasks
• Verify Task 2's LLM context only contains messages relevant to Task 2
• Verify iterations reset to 0 for each new task
• Verify single-task execution still works correctly

🤖 Generated with Claude Code