Code generation is one of the most exciting things in the world of Generative AI. Powerful models like Claude and Sonnet can understand instructions, break down complex tasks, and even build intricate systems on their own. This is very useful in code assistants like Cline.

However, sometimes Claude/Cline might have a hard time when a project gets too complicated. It might not have enough info or the ability to figure out what to do next. In these cases, the programmer has to step in, review the code, and guide Cline in the right direction. They also need to make sure the generated code is good and secure.

This raises an interesting question: can we get AI to help with this review and guidance process? Could we have an AI double-check the code made by Claude or another model, or even give suggestions when the main model gets stuck? To explore this, we can try a cool tool from Anthropic called MCP.

Read more in my Blog Post](https://airabbit.blog/)