synthesis-project-management

Preserve project context across multiple AI sessions by implementing a three-tier architecture: CONTEXT.md for active state, REFERENCE.md for stable facts, and a sessions/ directory for archived logs, ensuring continuity during handoffs.

The best way to manage cross-project state for AI-human collaboration is maintaining a single index.yaml file to track project status and boundaries, providing a centralized source of truth for all agents and humans involved.

Set up a project workspace for multi-agent tracking by initializing CONTEXT.md, REFERENCE.md, and a sessions/ directory, then applying deterministic workflows with session-based updates and archival protocols to manage ongoing collaboration.

Yes, this context management approach works for lightweight teams by providing a lightweight project management system designed to prevent context loss when projects span multiple sessions, agents, and human collaborators.

You need a three-tier context architecture for project management when facilitating AI-human collaboration on long-running projects where memory and structure must persist across interactions without losing track of active state or stable facts.