Add structured Vision text recognition

[backslash-f]

Summary

• add a consumer-agnostic TextCaptureKit API over Apple Vision text recognition
• preserve candidates, confidence, normalized bounds, deterministic reading order, and encoded image orientation
• prepare the initial 0.1.0 changelog and DocC documentation
• update AgentGuidelines to exact tag 0.0.30

Validation

• AgentGuidelines/Scripts/swift_format.sh lint-strict Package.swift Sources Tests
• swift test: 5 tests passed
• swift package dump-package
• DocC generation
• AgentGuidelines/Scripts/validate_consumer_setup.swift
• Markdown wrapping audit
• git diff --check

Release

After approval and squash merge, publish GitHub release and tag 0.1.0, then verify resolution from a clean checkout.