Summary

Adds MCP tool annotations (readOnlyHint, destructiveHint) to 121 tools across 6 MCP servers to help LLMs better understand tool behavior and make safer decisions about tool execution.

Servers Updated

Changes

• Added readOnlyHint: true to read-only tools (queries, fetches)
• Added destructiveHint: true to tools that create/update/delete data
• Added title annotations for human-readable display

Why This Matters

• Annotations provide semantic metadata that helps LLMs understand tool behavior
• LLMs can make better decisions about when to use tools and in what order
• Enables safer tool execution by distinguishing read-only from destructive operations
• readOnlyHint tells the LLM a tool is safe to call without side effects
• destructiveHint signals the LLM should be more careful before executing

Before/After Example

Before:

{
  name: "figma_get_file",
  description: "Get File Content - retrieve the full content of a Figma file."
  // No annotations
}

After:

{
  name: "figma_get_file",
  description: "Get File Content - retrieve the full content of a Figma file.",
  annotations: {
    title: "Get File",
    readOnlyHint: true
  }
}

Testing

• TypeScript servers build successfully (Intercom)
• Annotations appear in tool definitions
• Annotation values match actual tool behavior

🤖 Generated with Claude Code