Model Context Protocol (MCP) Interface Skill

This skill allows the agent to dynamically inspect and interact with any active Server-Sent Events (SSE) Model Context Protocol servers.

Available Sub-Scripts

All scripts are located in skills/mcp/scripts/ and should be executed using Python 3 relative to the workspace root.

---

🔍 1. list_tools.py

• Purpose: Discovers and prints all available tools and parameters schema from a target MCP server URL.
• Script Path: skills/mcp/scripts/list_tools.py
• Arguments:
  • --url [SSE_ENDPOINT_URL]: Target MCP server SSE endpoint URL (Required)
• Usage Example:

  python skills/mcp/scripts/list_tools.py --url https://mcp-server-wikipedia.up.railway.app/sse
  

---

⚡ 2. execute_tool.py

• Purpose: Executes a specific tool on a target MCP server with specified JSON arguments.
• Script Path: skills/mcp/scripts/execute_tool.py
• Arguments:
  • --url [SSE_ENDPOINT_URL]: Target MCP server SSE endpoint URL (Required)
  • --tool [TOOL_NAME]: Name of the tool to execute (Required)
  • --args [JSON_ARGUMENTS]: JSON string arguments representing tool parameters (Required)
• Usage Example:

  python skills/mcp/scripts/execute_tool.py \
    --url https://mcp-server-wikipedia.up.railway.app/sse \
    --tool search \
    --args "{\"query\": \"인공지능\"}"