feat(hex-ssh): Advanced transfer modes research backlog

[levnikolaevich]

Context

After the current hardening pass, the next level of transfer differentiation is no longer basic SFTP correctness. It is advanced transfer strategy.

These ideas are promising but should remain research backlog until the current engine is validated across real servers and the operational model is stable.

Why last

• advanced modes are easy to over-design before interop evidence exists
• resume and compression have real correctness and compatibility risks
• research should start from a stable baseline, not from open questions in the current transfer engine

What to evaluate

• stronger destination integrity verification beyond stat
  • checksum / hash verification
• guarded resume support
• adaptive compression for compressible payloads
• eventual delta-sync / rsync-like strategies

Constraints

• Do not weaken the current safe default path
• Do not adopt fastPut / fastGet as the unconditional default
• Treat resume and compression as opt-in until proven safe and useful
• Keep this issue research-only until at least #38 is complete

Definition of Done

• Each candidate mode has a written benefit/risk analysis
• Compatibility implications are documented
• API exposure is proposed only after the risk model is clear
• At least one candidate is either promoted to implementation issue or explicitly rejected

Related issues

• Should follow: #38
• Should not block: #37, #39, #40

Priority

Low-medium. Valuable for future differentiation, but not a prerequisite for current release quality.