An automated linter and schema-tuning assistant that scans other MCP servers to find and rewrite vague tool descriptions, ambiguous parameters, and missing structural constraints—instantly fixing the exact "tool smells" that cause AI agents to hallucinate, loop, or fail tasks.

Elevator pitch Automated linter that scans MCP servers, detects vague tool descriptions and ambiguous parameters, then rewrites schemas to eliminate AI agent hallucination and task failure.

Industry AI tooling & MCP infrastructure.

Problem

• Vague tool descriptions cause agents to misuse or skip available functions.
• Ambiguous parameter constraints lead to hallucination, infinite loops, and failed tasks.

Solution

• Scans live MCP servers and identifies "tool smells" in real time.
• Auto-generates precise descriptions, constraints, and structural fixes.

Tools

• SchemaScanner: Introspects MCP server tools, returns vagueness scores and ambiguity flags.
• DescriptionRewriter: Accepts tool metadata, returns clarified descriptions with examples.
• ConstraintInferrer: Analyzes parameter usage patterns, returns tightened enum/range constraints.
• SchemaValidator: Tests rewritten schemas against agent behavior logs, returns pass/fail report.

Widgets

• /dashboard: Real-time server health, linting results, and rewrite recommendations.
• /compare: Side-by-side before/after schema diffs with impact metrics.

Conversation starters

• "Scan my MCP server and show me which tool descriptions are confusing agents most."
• "Rewrite this parameter schema to eliminate ambiguity and add validation examples."