8agana/rusty-tools

rusty-tools

Latest tag

MCP (Model Context Protocol) server providing Rust development tools for AI assistants. Bridges the gap between AI code generation and Rust compiler validation.

Why rusty-tools?

AI assistants often struggle with Rust's strict compiler and complex error messages. rusty-tools provides direct access to Rust toolchain commands through MCP, enabling:

• Real-time validation - Check code without leaving the chat
• Error explanations - Understand cryptic Rust errors
• Code improvements - Get clippy suggestions instantly
• Format standardization - Apply rustfmt consistently

Installation

For Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "rusty-tools": {
      "command": "/path/to/rusty-tools/target/release/rusty-tools-server"
    }
  }
}

Building from source

git clone https://github.com/8agana/rusty-tools.git
cd rusty-tools
cargo build --release --bin rusty-tools-server

Releases

Download prebuilt binaries from GitHub Releases and verify the checksum.

# Download latest binary and checksum
curl -L -o rusty-tools-server \
  https://github.com/8agana/rusty-tools/releases/latest/download/rusty-tools-server
curl -L -o rusty-tools-server-SHA256.txt \
  https://github.com/8agana/rusty-tools/releases/latest/download/rusty-tools-server-SHA256.txt

# Verify checksum (macOS)
shasum -a 256 -c rusty-tools-server-SHA256.txt

# Make executable and install
chmod +x rusty-tools-server
sudo mv rusty-tools-server /usr/local/bin/

Then point your MCP client (e.g., Claude Desktop) to /usr/local/bin/rusty-tools-server.

Available Tools

Core Tools (Production Ready)

• cargo_build - Build project (produces artifacts)

  // Compiles the temp project and reports compiler output
  // Use `cargo_check` for faster type-checking without artifacts
  

• cargo_fmt - Format Rust code using rustfmt

  // Input: unformatted code
  // Output: properly formatted code
  

• cargo_clippy - Analyze code for improvements

  // Returns suggestions for better patterns, performance, and style
  

• cargo_check - Type-check without building

  // Fast validation of code correctness
  

• rustc_explain - Explain compiler error codes

  // Input: "E0308"
  // Output: Detailed explanation of type mismatch error
  

• cargo_fix - Automatically fix compiler warnings

  // Returns diffs of suggested fixes
  

Analysis Tools

• rust_analyzer - Deep code analysis and diagnostics (uses cargo check JSON)
• cargo_tree - Display dependency tree
• cargo_doc - Generate documentation

Security & Dependencies

• cargo_audit - Scan for security vulnerabilities
• cargo_search - Search crates.io for packages

Testing

• cargo_test - Run tests with visible output

  // Runs tests with --nocapture for full output visibility
  // Shows actual test results, not just pass/fail
  

Use Cases

For AI Assistants

• Validate generated Rust code instantly
• Understand compiler errors without context switching
• Apply formatting and best practices automatically

For Developers

• Get compiler feedback directly in AI chat
• Learn from error explanations
• Maintain code quality with integrated clippy

For Learning

• Understand Rust errors with clear explanations
• See formatting standards applied in real-time
• Learn best practices through clippy suggestions

Architecture

Built with rmcp 0.6.4, rusty-tools creates isolated Rust environments for each tool invocation, ensuring:

• No persistent state between calls
• Safe execution without side effects
• Fast response times

Limitations

Current limitations (by design for safety):

• No cargo run (binaries are never executed)
• Read-only operations on codebases

Output Format

All tool responses follow a consistent shape:

{ "status": 0, "success": true, "stdout": "...", "stderr": "", "duration_ms": 123 }

Use status/success for reliable checks; parse stdout for compiler output.

Contributing

Contributions welcome! Areas for improvement:

• Enhanced test output formatting
• Additional cargo subcommands
• Improved error message parsing

License

MIT

Author

Built by Sam Atagana (@8agana) - learned Rust in under 30 days to solve AI + Rust development friction.