project-memory

To persist project learnings across Claude Code sessions, you can store decisions and patterns using GitHub Issues as a searchable memory store. This records lessons after plan and review, binding decisions to specific issues to survive stateless session resets.

Tracking architectural decisions using GitHub Issues involves creating structured records with specific labels like type:decision and type:lesson. This ensures a consistent, searchable history of project patterns and failed approaches across your repository.

Storing project memory in GitHub Issues prevents context rot by applying tooling to query and consolidate memories. It uses a structured memory format defined in MEMORY.md with labeled comments to maintain consistent, searchable records for future sessions.

Yes, a MEMORY.md file is needed to define the structured memory format for recording lessons learned in GitHub Issues. It ensures consistent capture of successful patterns and failed approaches alongside labeled comments for searchable project history.

Yes, you can query and consolidate past project decisions bound to specific issues using the memory consolidation tooling. This retrieves architectural decisions and patterns through GitHub Issues labels to inform future development sessions.