The original title is "MCP Scorecard: 4-Layer Pre-flight Check Before You Publish a Server"

Short summary

The author released mcp-scorecard, a PyPI tool that runs four pre-flight checks on MCP servers before publishing, grading them A to F. It addresses passive token costs—descriptions, JSON schemas, and tool names sent to the model every turn—that runtime scanners miss. The tool ships as a CLI and as five MCP tools so Claude Code can audit servers in-session.

• •mcp-scorecard audits MCP server surfaces for token cost, description quality, naming, and leaks before publishing
• •A single verbose MCP server can silently burn 5,000+ tokens per turn without any tool being called
• •Ships as both a CLI and as MCP tools for in-session auditing within Claude Code

Generated with AI, which can make mistakes.