benjaminlevaillant-OnceUPapaTime/claude-company-brain

A practical architecture and guide to give Claude (Anthropic) a persistent, company-wide memory. It’s designed for teams and companies that use Claude in project work and want conversations to share context across users and sessions. What it is for: solve Claude’s lack of cross-session memory by layering short-term project journals, a curated long-term Company Brain, and a real‑time Data Lake. Who uses it: product/engineering/ops/CS teams and designated knowledge “Champions” who curate content; developers use a Git mirror for backups and integrations. Key features: - Three-layer memory: Journal (per-project short term), Company Brain (validated long term), Data Lake (live operational data) - Session tooling: /sync and /feed to capture conversation outputs - Governance: submission workflow, Champions, monthly “Brain Check”, changelog, sensitivity levels - Implementation: Notion as source of truth, Git mirror, Notion templates, freshness scoring, model-agnostic design and compliance/security guidance.

このRepoのdescription / GitHub topics / READMEのAI要約に出現したAI関連キーワードをカテゴリ別に表示。各バッジは該当カテゴリの詳細ランキングへリンクします。