docs: Add comprehensive agent-skills repository analysis

[daikh0]

Why

This PR documents a complete assessment of the agent-skills repository structure, including:

• Complete inventory of all 24 skills organized by lifecycle stage
• Framework analysis covering Markdown-based skill format, YAML frontmatter, and 3-tier testing system
• Templates and boilerplate for contributing new skills
• Platform support and installation methods

This analysis serves as a reference guide for understanding the repository architecture and onboarding new contributors.

What's included

REPO_ANALYSIS_VI.md (Vietnamese, 6.8 KB)

• Detailed breakdown of all 24 skills by lifecycle stage (Define, Plan, Build, Verify, Review, Ship, Meta)
• 8 slash commands with their invoked skills
• 4 agent personas (code-reviewer, test-engineer, security-auditor, web-performance-auditor)
• Template for creating new skills (folder structure, SKILL.md anatomy, eval case format)
• Step-by-step process for contributing new skills
• Complete framework documentation
• Installation options across 70+ platforms

COMPREHENSIVE_ANALYSIS.md (1.9 KB)

• Quick reference guide to repository structure
• Key technologies and frameworks used
• Installation methods
• Creating new skills summary

Key findings

• Total Skills: 24 (23 lifecycle + 1 meta-skill)
• Framework: Production-grade plugin system supporting 70+ agents
• Format: Markdown + YAML frontmatter (2-10 KB per skill)
• Testing: 3-tier validation system (Structural, Routing, Behavioral)
• Quality Bar: Specific, Verifiable, Battle-tested, Minimal

Framework components

• Skill format: Markdown + YAML frontmatter
• Command format: TOML + prompt templates
• Plugin system: plugin.json manifests
• Testing: 3-tier system (validate-skills.js, run-evals.js)
• Hooks: Bash + JSON for session lifecycle
• Platform support: 70+ agents (Claude Code, Cursor, Codex, Copilot, Cline, etc.)

Documentation

This analysis consolidates information from docs/skill-anatomy.md, CONTRIBUTING.md, README.md, and evals/README.md.