Restructure docs for useforge.ai sync pipeline

[initializ-mk]

Summary

• Reorganize flat docs/ directory into site-matching subdirectories: getting-started/, core-concepts/, security/, skills/, deployment/, reference/, plus faq.md
• Add YAML frontmatter (title, description, order) to every doc for automated sync
• Strip H1 headers, breadcrumbs, and navigation lines

Splits

• skills.md (27KB) → 6 files: skill-md-format, embedded-skills, writing-custom-skills, skills-cli, contributing-a-skill, your-first-skill
• deployment.md → 4 files: docker, kubernetes, production-checklist, monitoring
• configuration.md → 3 files: configuration, forge-yaml-schema, environment-variables
• security/overview.md → 3 files: overview, trust-model, audit-logging

New standalone pages

• reference/agent-skills-compatibility.md — compatibility matrix for skills across providers
• faq.md — frequently asked questions

Why

The useforge.ai sync pipeline fetches docs from this repo at build time. It expects:

1. Subdirectory structure matching the site's sections
2. YAML frontmatter with title, description, order
3. No H1 headers or navigation breadcrumbs (handled by the site layout)

Test plan

• Every new file has valid YAML frontmatter
• No content was lost — all original doc content exists in the restructured files
• No broken cross-references within the new structure