byterover

What This Skill Does

ByteRover is a sophisticated knowledge management system designed for OpenClaw AI agents to maintain long-term memory and project context. By utilizing the brv CLI, the agent creates a bridge between volatile session memory and persistent, file-based storage. It organizes project patterns, architectural decisions, and coding standards within a .brv/context-tree/ directory in your workspace. When triggered, ByteRover uses an integrated LLM to synthesize information from these Markdown files, ensuring the agent remains aligned with specific project requirements, avoiding hallucinations, and maintaining consistency across development sessions.

Installation

To integrate this skill into your environment, use the OpenClaw skill installer: clawhub install openclaw/skills/skills/byteroverinc/byterover. Additionally, ensure the underlying CLI is available on your system by running npm install -g byterover-cli. Once installed, configure the default LLM provider to enable intelligent retrieval and curation by executing: brv providers connect byterover. No external API keys are required for the default provider.

Use Cases

ByteRover is essential for complex, multi-day development tasks. It acts as an "architectural source of truth" for your codebase. Use it to store high-level design patterns, API interface definitions, and specific compliance rules that the agent needs to adhere to during task execution. It is particularly effective for teams where multiple agents might need to share a common knowledge base regarding project conventions or debugging steps discovered in previous sessions. By running brv query at the start of a session, the agent can instantly gain context without needing to parse the entire repository.

Example Prompts

1. "Before we start building the new user dashboard, check our stored project patterns using ByteRover to see how we handled layout components in previous views."
2. "We've finally decided to switch to Redis for caching. Please save this architectural decision to the context tree so we don't accidentally implement an in-memory store again."
3. "Summarize all the critical debugging steps and fixes we performed on the authentication module last week from the knowledge base."

Tips & Limitations

To maximize the utility of ByteRover, maintain a habit of using brv curate immediately following a significant breakthrough or decision-making process. The skill performs best when files are concise and Markdown-formatted. Remember that while brv query and brv curate work locally without authentication, features like cloud synchronization require additional setup. Do not use ByteRover for transient tasks like temporary debug logs or general knowledge that exists outside the scope of your current project, as this will clutter the knowledge base and potentially impact retrieval accuracy. Keep your context-tree organized to ensure the LLM provides relevant and specific synthesis.