feat(ai-agents): Add LM Studio integration and cybersecurity model testing

Implements AI agent testing with LM Studio and specialized cybersecurity models.

Features:
- LM Studio configuration and connection testing
- Support for 3 specialized cybersecurity models:
  - seneca-cybersecurity-llm (balanced performance)
  - seneca-cybersecurity-llm_x_qwen2.5-7b-cybersecurity (fastest & most detailed)
  - baronllm-v2-offensivesecurityllm (most thorough, slower)
- CLI integration with --config flag for AI agent configuration
- Comprehensive test suite for model comparison and validation
- Increased timeout support for larger models (5 minutes)
- Direct AI agent testing without full Cloud Safeguard run

Performance Results:
- Seneca Qwen2.5-7B: 24.76s response time, 2550 chars output
- Seneca Base: 28.17s response time, 2533 chars output
- Baron: 257.29s response time, 2336 chars output

Files Added:
- AI-AGENTS-TESTING.md: Complete testing guide
- configs/sim.ai-lm-studio.yaml: LM Studio configuration
- test-*.py: Comprehensive test suite for models and agents
- test-ai-agents.sh: Automated testing script

Files Modified:
- cmd/simctl/main.go: Added --config flag support
- agents/python/sdk/cloudsafeguard/llm/remote_backend.py: Increased timeout
- Various AI agent and parser improvements

Status: AI agents fully functional with LM Studio ✅