Add descriptions to skills

Author: webstackdev

skills/content_classification/SKILL.md

@@ -1,5 +1,6 @@
 ---
-name: content_classification
+name: content-classification
+description: Classify newsletter content into one supported editorial content type.
 input: title, content_text, url
 output: content_type, confidence, explanation
 ---

skills/deduplication/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: deduplication
+description: Decide whether two content items should be treated as duplicates.
 input: title, content_text, canonical_url, candidate_title, candidate_content_text, candidate_canonical_url, similarity_score
 output: is_duplicate, confidence, explanation
 ---

skills/entity_extraction/SKILL.md

@@ -1,5 +1,6 @@
 ---
-name: entity_extraction
+name: entity-extraction
+description: Extract tracked entity mentions and propose new entity candidates.
 input: title, content_text, project_id, tracked_entities
 output: mentions, candidate_entities, explanation
 ---

skills/relevance_scoring/SKILL.md

@@ -1,5 +1,6 @@
 ---
-name: relevance_scoring
+name: relevance-scoring
+description: Score how relevant a content item is for a project's newsletter topic.
 input: newsletter_topic, reference_similarity, title, content_text, url, source_plugin
 output: relevance_score, explanation, used_llm
 ---

skills/summarization/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: summarization
+description: Write a concise newsletter-ready summary for relevant content.
 input: title, content_text, newsletter_topic
 output: summary
 ---

skills/theme_detection/SKILL.md

@@ -1,5 +1,6 @@
 ---
-name: theme_detection
+name: theme-detection
+description: Generate one editor-facing theme suggestion from a high-velocity topic cluster.
 input: project_topic, cluster_context, recent_accepted_themes
 output: title, one_sentence_pitch, why_it_matters, suggested_angle
 ---