[FEATURE] Native /doctor command & Official Docs MCP to audit projects against latest best practices

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

The Claude Code team ships incredibly fast, with new features and recommended best practices dropping almost weekly. My biggest bottleneck right now is the manual research required to keep my projects up to date. Constantly reading new bangers from Boris & Team, watching multiple repos, and manually auditing my setups to align with your newest recommendations is a major pain.

Proposed Solution

If there were an official Claude Code Docs MCP, I could just tell my agent: "Audit this project against the newest Claude Code best practices and update my configurations." The agent would dynamically pull the living docs, instantly bridging the gap between what you shipped yesterday and what I'm building today.

Maybe you can integrate that update feature directly into Claude Code using a command.

Alternative Solutions

The current alternative is following your devs on X for updates, reading your docs weekly, and updating my projects manually. This doesn't make much sense when we literally have the tools (MCP) to automate this. I could build a custom MCP server myself and scrape your docs daily, but an official, maintained MCP from your team would be far more reliable and widely beneficial for the whole community.

Priority

Medium - Would be very helpful

Feature Category

MCP server integration

Use Case Example

Use Case 1: The Native /doctor or /audit Command

Let's say I open a Claude Code project I haven't touched in a month. In that time, the Anthropic team has shipped new ways to structure CLAUDE.md and handle tool routing.

Instead of reading a month of release notes on X or GitHub, I just open my terminal and type:
/doctor (or /audit)

Internally, Claude Code fetches the "living docs" via the official MCP, audits my current setup, and outputs a permission-based prompt:

"I noticed your CLAUDE.md is using legacy routing. Based on this week's update, the new best practice is [X]. I also see you aren't using the new pre-commit hooks feature. Would you like me to update your configuration to match the current recommendations?"

Use Case 2: Custom Agent Bootstrapping via MCP

If I'm building a new tool or using a custom agent workflow, I can just prompt my agent:

"Use the official Claude Docs MCP to audit my current setup and bootstrap the new subagent syntax that dropped this week."

The agent pulls the documentation directly from the MCP server, preventing it from hallucinating outdated CLI commands or legacy setup steps from its base training data.

Additional Context

_No response_