feat: implement AI agent framework and configuration analyzer

- Add AI agent base class with LLM backend abstraction
- Implement llama.cpp, MLX, remote, and API backends
- Add AI configuration analyzer agent with prompt engineering
- Update orchestrator to support AI agents with phased execution
- Add AI-specific configuration options and validation
- Update report generator to distinguish AI vs deterministic findings
- Add comprehensive AI agent documentation and examples
- All tests passing (21/21)