outcome: z.string().optional(),

**Client Information**
- LLM Client: [e.g. Claude Desktop, claude.ai, Cursor, VSCode]
- Client Config: [e.g. claude_desktop_config.json or screenshot of client configuration UI]
- MCP Server: [e.g. https://observability.mcp.cloudflare.com, https://docs.mcp.cloudflare.com]
- Prompt: [if applicable, add the prompt used]

**Describe the bug**
A clear and concise description of what the bug is. Any error logs are helpful.

**To Reproduce**
Steps to reproduce the behavior:
1. Connect to MCP Server...
2. Provide LLM prompt "..."
4. See error...

**Expected behavior**
A clear and concise description of what you expected to happen.

**Additional context**
Add any other context about the problem here.

Start by obtaining the missing client configuration, MCP server, prompt, reproduction steps, error logs, and expected behavior requested in the issue template. The current report names no files, tests, or entry points, so the work cannot be scoped or verified until those details are provided.

Written by the indexing model from the issue text.