Add comprehensive ShareHound OpenGraph support to BloodBash

[DotNetRussell]

This PR extends BloodBash with full offline analysis capabilities for ShareHound-generated BloodHound OpenGraph data. It allows red teamers and defenders to analyze large network share permission datasets without waiting for slow BloodHound CE ingestion.

Changes

• Parser & loader improvements: Enhanced load_json_dir() and build_graph() to robustly handle ShareHound OpenGraph JSON formats, including list-style exports, relationships/edges arrays, NetworkShareSMB, Folder, and File node types.
• New analysis module: Added print_sharehound_findings() with rich detection for:
• High-risk writable shares (CanWriteDacl, GenericAll, FullControl, CanDsWriteProperty, CanDsControlAccess, etc.)
• Shares accessible by “Authenticated Users” or “Everyone”
• Sensitive files on shares (VMDK, backups, databases, archives, etc.)
• FullControl principals
• Folder-level permissions
• Rich output: Added “Share Permission Summary” table and improved console reporting.
• Severity scoring: New finding categories (Share Write Access, Share Sensitive File) integrated with existing prioritization.
• CLI: New --sharehound flag + automatic inclusion when using --all.
• Testing: Added 4 new unit tests covering ShareHound findings, no-results case, Authenticated Users access, and Folder/permission table output. Full test suite (53 tests) passes.
• Documentation: Minor updates to README usage and feature list.

Why this matters
ShareHound can produce multi-GB OpenGraph JSON files on large enterprise networks. BloodHound CE ingestion can take days. BloodBash now provides fast, local, RAM-efficient analysis of exactly the same data.
Usage

python3 BloodBash /path/to/sharehound/output --sharehound --all
python3 BloodBash . --all --export=html

Testing
All existing SharpHound tests continue to pass. New ShareHound-specific tests were added and verified.