feat: Implement Lateral Movement Detection Agent

Implements comprehensive lateral movement detection agent for Phase 3:

Lateral Movement Agent Features:
- Detects unrestricted lateral movement paths between resources
- Identifies high-value pivot nodes with high connectivity
- Finds excessive network permissions and security group issues
- Analyzes network segmentation gaps and trust relationships
- Supports multiple vantage points (internal, external, IAM)
- Calculates attack path distances and trust chains

Core Components:
- agent.py: Main agent with vantage-aware analysis
- graph_analyzer.py: NetworkX-based graph analysis engine
- test_agent.py: Comprehensive test suite (11 tests)
- requirements.txt: Dependencies (networkx, cloudsafeguard-sdk)
- README.md: Complete documentation and usage guide

Graph Analysis Capabilities:
- Builds directed graph from resource nodes and edges
- Extracts network attributes (security groups, VPCs, subnets)
- Identifies network segments and trust relationships
- Calculates shortest paths and connectivity scores
- Detects overly permissive rules and public access

Agent Registry Integration:
- Registered in internal/orchestrator/agent_registry.go
- Supports internal, external, and IAM vantage points
- Enabled by default for security analysis

Testing Results:
- All 11 unit tests passing
- Successfully executed against mini-cloud example
- Generated 3 findings (pivot nodes with connectivity analysis)
- Proper error handling and vantage point filtering

Files Created:
- agents/python/examples/lateral_movement/agent.py
- agents/python/examples/lateral_movement/graph_analyzer.py
- agents/python/examples/lateral_movement/test_agent.py
- agents/python/examples/lateral_movement/requirements.txt
- agents/python/examples/lateral_movement/README.md

Files Modified:
- internal/orchestrator/agent_registry.go (added lateral_movement agent)

Status: ✅ Complete and tested
Next: Data Exfiltration Agent implementation