docs: reposition README around the production-operations direction (#900)

Context

Discussion #900 announced the project's direction: the official n8n MCP server is becoming the way workflows get written; n8n-mcp focuses on the production side of the lifecycle (validation, testing, deployment, diagnosis, multi-instance, coverage). The July 2026 competitive analysis documents the measurements behind this.

The README and docs still partially tell the old story. They should be brought in line with the announced positioning.

Tasks

• Audit README + docs for stale claims — in particular any remaining "n8n-mcp is N× cheaper per edit" framing, which is retired for current n8n versions (see the July analysis §3.1). Surviving, accurate claims: surgical in-field edits via patchNodeField (measured 5.3× on a ~1 KB Code node, scales with field size), strict validation, community-node coverage.
• Add a "Which MCP server should I use?" section near the top of the README: run both — the official server for drafting workflows on a current instance, n8n-mcp for honest validation, community/custom nodes, execution debugging, version history, and older n8n versions. Link Discussion #900 and the July analysis.
• Reorder the feature presentation to lead with the production-lifecycle tools (validation, versions, executions, autofix, audit) rather than node documentation.
• Add the maintainer consulting link/CTA.
• Pin Discussion #900 so it's the first thing visitors see.

Acceptance

A first-time visitor should understand within one screen: what this project is for today, how it relates to the official MCP server, and where the project is going.