Add MCP security best practices guide

## Feature Request: MCP Security Best Practices Cookbook Recipe

As MCP adoption grows, I think a security-focused cookbook recipe would be valuable for the community.

I've been building [mcp-safeguard](https://github.com/SyedAnas01/mcp-safeguard) — the first automated security scanner for MCP servers — and have identified the most common vulnerabilities in real-world MCP deployments.

### Proposed Cookbook Content
A new recipe covering:
1. **Prompt injection defense** — how to sanitize tool outputs before returning to AI context
2. **Credential hygiene** — safe handling of API keys and tokens in MCP servers
3. **Network security** — preventing SSRF through URL validation
4. **Tool poisoning prevention** — validating tool definitions at load time
5. **Security scanning integration** — adding mcp-safeguard to CI/CD pipelines

### Why This Matters
- 500+ MCP servers now published on GitHub and mcpservers.org
- Most authors are not security experts
- MCP-specific attack vectors differ significantly from traditional web app vulnerabilities

I'd be happy to write this recipe and submit a PR. Would the team be interested?

Reference: https://github.com/SyedAnas01/mcp-safeguard