kortix-memory

To preserve project context across AI sessions, you can use a structured project memory system that stores durable knowledge like architecture decisions and conventions in version-controlled files. This ensures team members and AI sessions access consistent historical context without re-debating solved problems.

The best way to manage team knowledge involves using a standardized memory file structure with an index and topic-specific sub-files. This approach curates durable project information such as purpose, integrations, and workarounds, keeping knowledge organized and easy to navigate for all contributors.

To safely update shared project memory files, you should follow a reviewed change request workflow to land edits to the main branch. This end-to-end edit process ensures that any additions, updates, or reorganizations to the memory files are safely reviewed before integration.

Version-controlled project memory should include durable knowledge such as purpose, architecture decisions, integrations, conventions, and workarounds. A structured memory rubric provides clear guidelines on what belongs in memory and what should be discarded to eliminate lost context.

Yes, this project memory system is specifically designed for Kortix-using software development teams. It provides a shared, version-controlled project brain accessible across all AI sessions and contributors, satisfying requirements for a standardized memory file structure and reviewed change request workflow.

To document a new branching convention for future AI sessions, you add the decision to the project's conventions memory file. This stores the convention in a durable, version-controlled format so no future team member or AI session misses the established standard.