talk-stage3-concepts

Extracting concepts from a talk summary is done by loading the markdown summary and optional timeline to identify ideas, categorize them by domain, and output a numbered catalogue with talk-potential scores.

Scoring and categorizing talk points involves assigning a domain category and a HIGH, MEDIUM, or LOW talk-potential rating to each extracted concept, producing a structured catalogue for clear mapping.

Concept enrichment uses the provided repository path to generate an enriched catalogue version, adding per-concept metadata alongside the main output file for deeper context.

A timeline file is optional for generating a concept catalogue; the extractor primarily requires the summary file, but loading the timeline enriches the concept scoring process.

The concept extraction output uses a markdown file format, specifically named talks/{YYYY}-{slug}-concepts.md, to store the numbered and categorized concept catalogue.

Automating concept categorization is fully supported; the pipeline loads source materials, assigns domain categories, applies talk-potential scores, and generates the catalogue automatically.