Proposal: persistent project memory for repository Q&A agents

Hi, I’m Vivek Gupta, Founder & CEO at [MemCode](https://memcode.in). Petercat helps teams create conversational Q&A agents around GitHub repositories, and those agents become much more useful when they can remember project-specific questions, accepted answers, maintainer preferences, and user context across sessions.

Would Petercat consider an optional MemCode memory integration? MemCode could provide a scoped memory layer for repository, team, and user context, exposed through a small tool or service adapter. Retrieval could stay separate from the repository knowledge base, with clear provenance, opt-in writes, configurable retention, and explicit forget/delete controls.

I’d be happy to put together a minimal integration example or PR after learning how Petercat prefers to structure agent state. A short technical call or a draft PR is easy on my side—does this seem useful for the project?