# Arscontexta plugin for Claude Code Arscontexta is a [[Claude Code]] plugin that automatically generates personalized knowledge management systems through conversation. It creates a persistent thinking infrastructure — a "second brain for your agent" — so Claude remembers your knowledge structure, methodologies, and preferences between sessions instead of starting blank each time. The name references "Revelatio memoriae" (revelation of memory), framing the project within a historical tradition of external thinking systems — from Llull's combinatorial wheels to Bruno's memory theaters — now executable by LLMs. ## How It Works A 20-minute conversational setup where you answer 2-4 questions about your domain. The engine then generates a complete Markdown-based vault tailored to your needs: - **`self/`**: Agent identity and methodology - **`notes/`**: The actual knowledge graph - **`ops/`**: Operational tracking ## Key Features - **Research-backed design**: 249 interconnected research claims from cognitive science, Zettelkasten, Cornell Note-Taking, and agent architecture principles - **Processing pipeline**: Six phases — Record, Reduce, Reflect, Reweave, Verify, Rethink — each spawning fresh subagents to prevent context window degradation - **Automation hooks**: Session orientation, write validation, auto-commit, and session capture - **Domain-native commands**: Generated skills use vocabulary matching your field (e.g., `/claim` instead of `/note` for researchers) - **Maps of Content**: Hierarchical navigation at hub, domain, and topic levels - **Semantic search**: Optional qmd integration for concept matching - **Multi-domain support**: Expand systems to cover new knowledge areas post-setup ## References - GitHub: https://github.com/agenticnotetaking/arscontexta - Website: https://www.arscontexta.org ## Related - [[Claude Code]] - [[Knowledge Management (KM)]] - [[AI Agents]] - [[Personal AI Infrastructure]] - [[Compound Engineering Plugin]] - [[Zettelkasten method]]