use_aws_mcp

The use_aws_mcp server provides AWS CLI functionality through a standardized MCP interface. It exposes a single tool named use_aws that lets MCP clients execute AWS CLI operations with automatic parameter formatting, safety checks, and user agent handling. Clients can craft requests that specify the AWS service, operation, region, and optional parameters and profile, and the server will translate those into proper AWS CLI invocations while enforcing read-only vs write operation semantics. This makes AWS interactions portable across different MCP-enabled tools without requiring direct shell access to the AWS CLI from the client environment.

To use the server, start it through your MCP client configuration and issue JSON-RPC style requests to the use_aws tool. For example, you can ask the server to list S3 buckets in a region, or describe EC2 instances with specific filters. The server will handle parameter kebab-casing, determine whether an operation is read-only or write, and return structured results suitable for downstream tools to present to users. If needed, you can provide a specific AWS profile and region to target a particular account or environment.

Prerequisites:

• Rust (1.70 or later) and Cargo installed on your machine
• AWS CLI installed and configured, or equivalent AWS credentials access
• AWS credentials available (via AWS CLI config, environment variables, or IAM roles)

Installation steps:

1. Install Rust and Cargo if you haven't already:
   • macOS/Linux: curl https://sh.rustup.rs -sSf | sh
   • Windows: follow rustup installation instructions from https://rustup.rs/
2. Build the MCP server in release mode:

   cargo build --release
   

3. Run the MCP server locally (the binary will be at the following path):

   ./target/release/use_aws_mcp
   

4. Confirm the server is running by sending a test request via your MCP client or observing logs. The server communicates via stdin/stdout using JSON-RPC protocol.
5. (Optional) If you want to test with the example tool description, build examples or run cargo run --example description_demo as described in the README.

Tips and considerations:

• Ensure your AWS credentials are configured and accessible to the process running the MCP server. If you are using non-shell MCP clients, explicitly set the desired AWS profile using the profile_name parameter in use_aws requests.
• The server supports read-only operation detection based on operation name prefixes and will request explicit acceptance for write operations.
• Outputs may be truncated to protect memory; large results are capped to a reasonable size.
• If you encounter AWS CLI or credential issues, verify that AWS CLI is installed, reachable in PATH, and that the configured profile/credentials are valid for the requested operations.
• Use the provided JSON schema for input to ensure proper parameter formatting and kebab-case conversion for CLI compatibility.