What problem does it solve? AI coding agents lose all project knowledge between sessions, forcing developers to repeatedly re-explain architecture decisions, coding patterns, and configuration details. This Skill provides a persistent, version-controlled memory layer so agents can store and retrieve project context across sessions, machines, and teammates. ## Core Features & Use Cases - Knowledge Curation and Querying: Record design decisions, business logic, and configuration patterns with brv curate, then retrieve them with semantic search via brv query. - Version Control for Context: Branch, commit, merge, push, and pull the context tree with Git-like commands (brv vc), enabling team collaboration on shared project knowledge. - MCP and Multi-Agent Integration: Expose memory tools to Claude Desktop and 22+ AI coding agents through the Model Context Protocol server (brv mcp), with support for 20+ LLM providers. - Use Case: A team curates API design decisions and database configuration into the context tree, pushes it to ByteRover Cloud, and a newly onboarded developer clones the shared space and immediately queries setup instructions and endpoint conventions. ## Quick Start Install ByteRover CLI with npm install -g byterover-cli, then ask the agent to initialize ByteRover in the current project and curate a key architectural decision into the context tree.