MCP Linter -- Validate MCP Tool Definitions

mcp_manifest_lint is a free MCP tool definition linter that validates your tools for anti-patterns, missing descriptions, bad parameter schemas, and naming issues. It is the only linter purpose-built for MCP tool definitions. No API key required.

Poor tool definitions cause agents to pick the wrong tool, pass bad parameters, or skip tools entirely. The MCP linter catches these problems before they reach production.

What It Checks

The linter runs every tool definition through these specific checks:

• Missing or inadequate descriptions -- tools without descriptions, or with descriptions too vague for an agent to understand when to use the tool
• Bad parameter schemas -- missing type fields, parameters without descriptions, no constraints on string/number inputs, missing required arrays
• Naming anti-patterns -- inconsistent casing (mixing camelCase and snake_case), ambiguous names like do_stuff or handle_request
• Missing annotations -- no readOnlyHint, destructiveHint, or idempotentHint, which agents use to assess risk before calling a tool
• Overly broad parameters -- inputSchema with type: "object" but no properties defined, or parameters that accept any string without guidance
• Agent confusion risks -- tool definitions where an agent cannot distinguish this tool from another, or cannot infer correct usage from the definition alone

Each issue is reported as an error (must fix) or warning (should fix), with a specific description of what is wrong and how to fix it.

How to Use It

Connect to the Agentic Platform MCP server, then call mcp_manifest_lint with your tool definitions as a JSON array in the tools_json parameter.

Input

mcp_manifest_lint({
  tools_json: '[{"name": "do_stuff", "description": "", "inputSchema": {"type": "object"}}]'
})

Output

MCP Manifest Lint Report
========================

Tool: do_stuff
  FAIL  Missing description
  FAIL  Vague tool name -- agents cannot infer purpose
  WARN  No inputSchema properties defined
  WARN  Missing annotations (readOnlyHint, destructiveHint)

1 tool checked. 2 errors, 2 warnings.

You can pass multiple tools at once. The linter checks each tool independently and returns a combined report.

Connect

claude mcp add agentic-platform --transport http https://api.asalvocreative.com/mcp

{
  "mcpServers": {
    "agentic-platform": {
      "url": "https://api.asalvocreative.com/mcp"
    }
  }
}

npx -y @smithery/cli@latest install agentic-platform--andysalvo.run.tools --client claude

Why Lint MCP Tool Definitions?

MCP tool definitions are the interface between your server and every AI agent that connects to it. A bad definition means agents will misuse your tool, ignore it, or call it with wrong parameters. Linting catches these issues at development time.

Common problems the linter finds: tools with no description that agents skip entirely, parameter schemas missing types so agents guess wrong, and tool names so vague that agents pick a different tool instead.

About the Agentic Platform

The Agentic Platform is an MCP server that provides free agent diagnostics and expert skill files. In addition to the MCP linter, it offers an agent health check, cost estimator, and service trust evaluator. All tools are free with no API key required.