Hey, thanks for the feature request. The idea of an agent that remembers the context of past changes and understands the why behind the code that’s already there makes sense and fits well.
A couple things that already partly cover this, and that might help you right now:
- Memories: the agent can save facts and decisions across sessions, so you don’t have to re-explain context every time.
- Search across past chats: the agent can search previous conversations to pull up context from earlier work.
- Rules
.cursor/rules: for persistent project guidance, so the agent can consistently keep architectural decisions and agreements in context: Rules | Cursor Docs - Plus codebase indexing and semantic search help the agent understand the repo while it works.
We don’t have an automatic memory graph on top of change history with Origin integration yet. I’ll leave this as a feature request so it can gather votes from the community. Agent memory is being actively discussed, here are a couple related threads if you want to add your vote or more context:
- [MCP] Add Persistent Memory in Cursor
- Persistent Agent Knowledge Bases Turn Subagents into Long-Lived Engineering Teammates
Let me know if any of the above doesn’t cover your use case. I’d like to understand exactly where it falls short.