project-context

To synchronize project state across multiple AI agents, use structured memory files like project-state.md and sprint-context.md to load session context and track ongoing work.

The best way to record architecture decisions for a team of agents is to log Architecture Decision Records in a dedicated decisions.md file, ensuring synchronized context.

You can generate an agent handoff summary with sprint context by applying structured memory read and write protocols to capture ticket updates and current sprint goals.

Yes, you can enforce strict versioning and safe write guards for multi-agent memory updates to prevent conflicts when agents read, update, or share project state.

Agents need to load sprint context before starting development work to ensure they have the latest ticket statuses, preventing duplicated effort and maintaining team synchronization.

You should not use a shared memory directory for project state if your workflow lacks multi-agent collaboration or requires isolated execution without synchronized architecture decisions.