Luma Core Topic Selection

There is no topic reviewer. Select locally.

Inputs

• 01_profile.json
• optional 01_profile_extra.md
• 02_content_history.json
• 02_raw_signals.json

Procedure

1. Cluster related social and web signals into candidate themes.
2. Remove candidates duplicating historical title, thesis, or angle.
3. Score remaining candidates on audience relevance, creator stance, freshness, conflict, material availability, and ability to support a full argument.
4. Compare at least three candidates.
5. Select exactly one and record rejection reasons for at least two alternatives.