docs: rewrite all 34 MCP tool descriptions for AI agent consumption

MCP tools are called by AI agents (Claude, GPT) — the description is
the only context an agent has to decide WHEN to use a tool. Previous
descriptions were too terse and used internal jargon.

New pattern: What it does → When to use it → What it returns

Examples:
- "Find nodes that consume from a target (CONSUMES/LISTENS edges)"
  → "Find all services, handlers, or functions that consume/listen
     from a given topic, queue, or event source. Use when asked
     'what reads from this topic?'. Returns consumer nodes with
     their kind, label, and file location."

Also updated 17 @McpToolParam descriptions with full enum values
and default/max hints.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: aksOps