What problem does it solve? Interacting with Model Context Protocol servers normally requires writing client code or using a dedicated host application, making quick exploration and ad-hoc tool calls cumbersome. This Skill provides a direct command-line interface to list servers, inspect tool schemas, and execute MCP tools without boilerplate. ## Core Features & Use Cases - Server and Tool Discovery: List all configured MCP servers and their tools, search tools by name with glob patterns, and optionally include descriptions. - Schema Inspection: Retrieve the full JSON input schema for any tool before calling it, so arguments are always correctly structured. - Tool Execution with Scripting Support: Call tools with JSON arguments via inline strings, heredocs, or piped input, and get JSON or raw output for chaining with jq and shell pipelines. - Use Case: A developer wants to read a file through the filesystem MCP server: they run mcp-cli filesystem/read_file '{"path": "./README.md"}' and immediately get the result, then pipe JSON output into jq for further processing. ## Quick Start Ask the AI to list all available MCP servers and tools by running mcp-cli, then call a specific tool such as reading a file through the filesystem server.
