What problem does it solve? MCP server tools often ship with unclear descriptions, weak schemas, and untested error paths, making it hard to know whether they will behave correctly in a Claude Code session. This Skill provides a systematic six-phase workflow to audit, test, and rate every MCP tool available in the current session. ## Core Features & Use Cases - Tool Discovery & Inventory: Lists all tools with the mcp__ prefix, groups them by server, and reports parameter counts and description previews. - Quality Analysis: Scores naming conventions, description quality, and parameter efficiency with severity indicators (critical, warning, suggestion). - AAA Test Generation & Execution: Creates Arrange-Act-Assert test cases covering happy paths, invalid inputs, and edge cases, with mandatory confirmation before running mutating tools. - Cross-Tool Analysis: Detects redundant tools, consolidation opportunities, and missing CRUD capabilities across servers. - Use Case: After adding a new MCP server, ask for a full audit to receive a report rating each tool, flagging a silent failure on wrong-type input, and recommending enum constraints for free-text parameters. ## Quick Start Ask the assistant to test all MCP tools in the current session and generate a full quality report.
