Add video-streaming category schema

[keepsloading]

Goal

Add a video-streaming category schema for apps that understand watch preferences.

This should help Memact convert messy viewing activity into user-readable Wiki entries that users can approve, edit, reject, or delete.

Context examples

Video apps may know things like:

• favorite genres
• disliked genres
• watch frequency
• preferred content length
• repeated creators or channels
• skipped content types
• language/subtitle preferences
• family/shared-device ambiguity

Work to do

• Define useful context fields for the video-streaming category
• Add example raw app context dumps
• Add normalized context examples
• Add user-facing Wiki entry templates
• Mark sensitive or risky fields clearly
• Suggest category-level permissions apps should request
• Add tests for safe context shaping

Important product rule

Do not assume every watch event is a stable user interest. Curiosity clicks and shared-device activity should be treated carefully.

Acceptance criteria

• Schema distinguishes strong preferences from weak signals
• Raw app input examples exist
• Normalized output examples exist
• Wiki entry examples are human-readable
• Tests cover at least one valid and one unsafe/low-confidence case

[keepsloading]

Maintainer note for SSoC26 contributors:

Please comment before starting so work does not get duplicated. Keep the first PR small and focused.

For category schemas, apps should propose context; users control what becomes accepted memory. Prefer user-readable summaries over raw personal data, avoid sensitive trait inference, and do not write fake certainty.

Before opening a PR, read MEMACT.md and make sure your schema includes examples, Wiki entry outputs, permission suggestions, and at least basic tests.