Empowers Claude Code with persistent memory and an evolving personality to act as a consistent, long-term AI teammate.
Team Memory is a sophisticated framework designed to transform Claude Code from a stateless assistant into a durable teammate by maintaining context across sessions. It utilizes a structured memory hierarchy consisting of Core, Archival, and Journaling tiers to store project decisions, coding patterns, and user preferences. By leveraging automated session-end consolidation and confidence-based pruning, the skill ensures that the AI assistant evolves its personality and knowledge base over time, significantly reducing the need for repetitive onboarding and manual context sharing.
主な機能
01Support for multiple teammate personas with switchable identities and skills.
02Confidence-based memory aging to maintain high-quality, relevant information.
03Seamless CLI integration for initializing, listing, and auditing memory blocks.
041 GitHub stars
05Multi-tiered memory storage (Core, Archival, Recall) for optimized context loading.
06Automated session consolidation that extracts insights and prunes stale data.
ユースケース
01Retaining long-term architectural decisions and project-specific patterns across weeks of development.
02Developing specialized AI personas tailored to specific tech stacks or team roles.
03Automating the capture of debugging discoveries and 'lessons learned' for future reference.
