This research introduces a novel approach to preserve software architecture integrity through prompt-engineered small language models that identify architectural tactics in code.

• Addresses the critical problem of architectural erosion where developers inadvertently remove or alter architectural decisions
• Leverages small language models with specialized prompt engineering for improved detection accuracy
• Focuses on maintaining non-functional requirements by making architectural knowledge more explicit
• Provides developers with tools to preserve architectural integrity during code modifications

For engineering teams, this approach offers a practical way to protect architectural decisions that support system quality attributes, potentially reducing maintenance costs and technical debt.

Enhancing the Accuracy and Comprehensibility in Architectural Tactics Detection via Small Model-Augmented Prompt Engineering