docs(spec): add ClearCam document scanner design

Thin shell over Google ML Kit Document Scanner (capture, edge
detection, perspective correction, filters, multi-page, JPEG/PDF)
plus ML Kit Text Recognition (Korean) for OCR. No in-app library;
scan then share/save. Play Store target, Play Services allowed.

Constraint: no CAMERA permission (scanner runs in Play Services process); files stay under 300 lines.
Confidence: high
Scope-risk: local
Reversibility: trivial
Directive: next produce implementation plan via writing-plans.
Tested: design self-review (no placeholders/contradictions).
Not-tested: no code yet.
Related: ClearCam greenfield

🐙 Autopus <noreply@autopus.co>