vibecheck

VibeCheck is a high-performance Zig-based CLI tool that scans your codebase for unfinished vibes like To-Dos, hardcoded secrets, and debug prints. When used with the MCP (Model Context Protocol) mode, VibeCheck exposes its capabilities as an MCP server so AI agents can interact with and orchestrate analyses through the standard MCP interface. To enable MCP mode, run the binary with the --mcp flag. The tool itself emphasizes fast, machine-parseable output (JSON) suitable for CI/CD pipelines and integration with automation tools. You can also use it in regular human-readable mode for quick, manual scans, or in CI/CD mode with --json for structured results and exit codes that reflect scan outcomes. Additionally, VibeCheck supports custom vibe packs and GitHub Actions annotations for streamlined automation in large repos.

Core commands of interest when using MCP:

• vibecheck --mcp: Start the MCP server interface for AI agents.
• vibecheck . --json: Run a full scan in machine-readable JSON format.
• vibecheck list-patterns: Enumerate active and available patterns.

As an MCP server, you can query or control analyses via the MCP protocol, enabling automated reasoning, orchestration, and integration with other tools in your software lifecycle.

Prerequisites:

• A working Zig toolchain (Zig 0.13+)
• Git

Installation steps:

1. Clone the repository: git clone https://github.com/copyleftdev/vibecheck.git
2. Navigate into the project: cd vibecheck
3. Build the project with release optimizations: zig build -Doptimize=ReleaseSmall
4. The compiled binary will be available at: zig-out/bin/vibecheck
5. Run the MCP server: ./vibecheck --mcp

Notes:

• The binary is platform-specific; ensure you build on the target OS/architecture.
• Optionally, you can run with --json to get machine-friendly output or with --github for CI/CD integration.

Tips and common issues:

• If you encounter missing dependencies during build, ensure Zig 0.13+ is installed and correctly configured
• Use the provided configuration file vibecheck.toml to customize ignores, thresholds, and external packs
• For MCP usage, ensure the agent communicating via MCP adheres to the standard MCP request/response format
• The MCP server supports JSON output, making it easy to parse in automation pipelines
• If you need to customize patterns, explore the extra_vibes.json in the examples directory and reference it via the packs setting in vibecheck.toml