As computer science grows in popularity, more students are enrolling in introductory CS courses (CS1), many of whom are new to the field and lack fundamental skills such as debugging, critical thinking, and code design. This leads to a reliance on undergraduate peer mentors (PMs) or graduate teaching assistants (TAs) during office hours. However, students often exhibit ineffective help-seeking behaviors, such as directly asking for answers, which makes it difficult for PMs and TAs to provide meaningful guidance. Consequently, many mentors unknowingly adopt practices that do not align with their goals of fostering student learning and skill development. To address this, we developed a structured reflection tool called Q&A Buddy hosted on Google Sheets. The tool scaffolds reflection to help PMs:

1. Recognize misalignments between their mentoring goals, actions, and underlying beliefs.
2. Explore alternative actions that better support student learning.
3. Motivate them to adopt these actions in future office hours.

Our findings show that the tool is effective in helping PMs identify these misalignments and understand the need for alternative beliefs. However, translating this reflection into actionable change remains a challenge. This quarter, we focused on addressing this by integrating considerations of real-world constraints into the reflection process. By creating space for PMs to identify and strategize ways to overcome these constraints, we aim to make alternative actions feel more feasible and actionable. Early user tests suggest that this acknowledging constraints approach is starting to show some improvement in how PMs perceive the "do-ability" of alternative strategies. However, the problem is not yet fully resolved, and we continue to explore ways to improve this process as a key next step.