feat: add --interactive flag to run command for review mode

[davidturnbull]

Overview

• Adds --interactive to run to review and edit AI-generated translations before writing.

Behavior

• Presents a colorized diff of current vs proposed translations per file/locale.
• Lets you approve all changes, skip them, or edit values individually in your editor.
• Serializes processing for a clear, deterministic review flow.
• Skips incremental writes; only writes after review approval.
• Works with existing filters and flags (e.g., --bucket, --file, --target-locale, --force).

Usage

• lingo.dev run --interactive
• lingo.dev run --interactive --bucket json --target-locale fr-FR

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.