feat: add auditable workflow recovery commands

[MageByte-Zero]

Summary

• add deterministic recovery context resolution and auditable human/JSON summaries
• add ssf resume, ssf switch, and ssf save with checkpoint-compatible write semantics
• add /ssf:resume, /ssf:switch, and /ssf:save command assets
• distribute and validate the recovery command tree for WorkBuddy
• update user docs, state-machine guidance, artifact contract, changelog, and release checks

Why

Issue #47 exposed a recovery gap: the repository already stored workflow, checkpoint, handoff, and SDD execution evidence, but users had no unified commands to inspect or resume that context. Slash-command adapters and WorkBuddy distribution also did not expose the recovery flow.

The implementation keeps the existing eight-state model. Resume and switch are read-only context recovery operations; save reuses the existing checkpoint contract and performs no partial writes on invalid input.

Closes #47

Validation

• npm run build
• focused recovery, command-asset, execution-control-plane, and WorkBuddy tests: 59/59 passed
• Node 20 full suite: 474/474 passed
• artifact validation: 0 errors, 0 warnings
• skill lint: 0 errors
• git diff --check

Known baseline

On Node 24, six existing guard tests with fixed 5000 ms timeouts can exceed the limit only in the full concurrent suite. Their focused runs pass, and the complete Node 20 support baseline passes. No out-of-scope timeout changes are included here.