Worked on improving the reliability of the plan lifecycle in the agentscope-ai/agentscope repository by addressing a bug related to plan change hooks during historical plan recovery. Focused on backend development and testing using Python, the work ensured that hooks are consistently triggered when restoring historical plans, closing a gap where plan state changes were not properly signaled. Automated tests were added to verify that downstream triggers fire as expected after restoration, strengthening the signaling mechanism for plan state changes. This targeted fix enhanced the robustness of the system’s event handling, contributing to more predictable and maintainable backend workflows.