What problem does it solve? When multiple trending-topic candidates compete for editorial attention, editors need a consistent way to adjust for content saturation and reader relevance before final topic selection. This Skill compares all temporary exploration cards and produces structured competitive corrections without computing final scores itself. ## Core Features & Use Cases - Competitive Correction Output: Returns strict JSON with per-candidate saturationPenalty (0-15), readerStakeScore (0-5), and a short reason for each candidate. - Duplicate Event Detection: Raises saturation penalties when the same event reappears under different subjects, dimensions, or titles, and prevents combination recommendations from selecting substantially identical angles. - Meta-Narratives and Combinations: Produces up to 3 meta-narratives plus a primary/stable/dark-horse combination recommendation for the topic-selection re-ranking stage. - Use Case: In a WeChat content pipeline, after hotspot exploration cards are generated, run this stage to re-rank candidates by penalizing saturated angles and surfacing a differentiated topic combination before article production. ## Quick Start Ask the agent to run the hotspot synthesis stage over the current batch of exploration cards and return the JSON corrections with saturation penalties, reader stake scores, and a recommended topic combination.
