byterover-headless

What This Skill Does

The byterover-headless skill enables OpenClaw AI to interact directly with the ByteRover CLI, providing a robust interface for managing project-specific knowledge bases. It allows the AI to query architectural decisions, retrieve implementation patterns, and curate new context into the ByteRover context tree. By leveraging the --headless --format json flags, this skill ensures that all interaction is machine-parseable and suitable for automated workflows, making it an essential tool for maintaining consistency across complex codebases.

Installation

To integrate this skill into your environment, use the OpenClaw Hub CLI:

clawhub install openclaw/skills/skills/byteroverinc/byterover-headless

Once installed, ensure you have initialized your ByteRover project using the brv init command within your project root. You must provide a valid API key obtained from your ByteRover settings, along with the correct team and space identifiers to ensure proper context scoping.

Use Cases

• Automated Documentation: Automatically curate new design decisions into the context tree during pull request reviews or architectural refactoring sessions.
• Knowledge Retrieval: Query the codebase for complex implementation details, such as how authentication middleware handles JWT tokens, without manually searching through thousands of lines of code.
• Context Synchronization: Use push and pull commands to sync local knowledge bases across team environments, ensuring every agent has access to the most recent project patterns.

Example Prompts

1. "Check the current project status and confirm if ByteRover is initialized; if not, ask me for my API key and the necessary team/space details."
2. "Query the knowledge base to explain the current implementation strategy for our database connection pooling."
3. "Curate this decision: We have moved from REST to GraphQL for the primary user profile endpoint. Please include the changes from src/api/user.ts in the context."

Tips & Limitations

• Efficiency: Always use the --headless flag. Without it, the CLI may prompt for interactive input, causing the agent to stall.
• Context Limits: When using brv curate with the --files flag, you are limited to a maximum of 5 files per command. Break larger context updates into multiple steps to stay within this boundary.
• Authentication: Keep your API keys secure. Use environment variables if possible, and ensure the agent only has access to relevant workspaces.