Add self-study-agent skill with feedback loop

- SKILL.md: Main skill definition with 20-iteration feedback loop
- config.yaml: Pipeline parameters (loop, sources, questions, adaptation)
- pipeline/: 5-step learning pipeline (topic selection → source discovery → extraction → synthesis → report)
- feedback/: 4 feedback mechanisms (self-testing, spaced repetition, interleaving, adaptation)
- Learning science principles: active recall, spaced repetition, desirable difficulty, interleaving, delayed feedback

skills/self-study-agent/config.yaml

@@ -0,0 +1,88 @@
+# Self-Study Agent - 파이프라인 파라미터
+
+# 루프 설정
+loop:
+  max_iterations: 20
+  early_exit_accuracy: 0.95
+
+# 주제 선택
+topic_selection:
+  strategy: "balanced"
+  user_goals: []
+
+# 소스 발견
+sources:
+  channels:
+    web_search: true
+    academic: true
+    github: true
+    documentation: true
+    local_files: false
+  max_per_channel: 5
+  max_total: 10
+
+# 추출
+extraction:
+  model: "small"
+  max_sources: 5
+  cross_reference: true
+  min_sources_for_fact: 2
+
+# 문제 생성
+question_generation:
+  # 문제 유형 비중
+  type_distribution:
+    direct_recall: 0.4      # 직접 회상
+    application: 0.3        # 적용 질문
+    variation: 0.2          # 변형 질문
+    connection: 0.1         # 연결 질문
+
+  # 난이도 분포
+  difficulty_distribution:
+    easy: 0.3
+    medium: 0.4
+    hard: 0.3
+
+  # 세션당 문제 수
+  questions_per_session: 10
+
+# 피드백
+feedback:
+  # 피드백 생성 규칙
+  include_explanation: true
+  include_related_concepts: true
+  include_source_reference: true
+  include_suggested_review: true
+
+# 간격 반복 (FSRS)
+spaced_repetition:
+  algorithm: "FSRS"
+  fsrs_params: [0.4, 0.6, 2.4, 5.8, 4.93, 0.94, 0.86, 0.01, 1.49, 0.14, 0.94, 2.18, 0.05, 0.34, 1.26, 0.29, 2.61]
+  min_interval_days: 1
+  max_interval_days: 365
+  target_retention: 0.9
+
+# 교차 학습
+interleaving:
+  enabled: true
+  mix_previous_sessions: 3   # 이전 세션 주제 포함 수
+  mix_ratio: 0.3             # 이전 주제 비중
+
+# 적응
+adaptation:
+  # 정답률 기반 적응
+  low_accuracy_threshold: 0.5
+  high_accuracy_threshold: 0.9
+  low_accuracy_consecutive: 3
+  high_accuracy_consecutive: 5
+
+  # 난이도 조절
+  difficulty_adjustment:
+    enabled: true
+    min_difficulty: 0.1
+    max_difficulty: 0.9
+
+  # 문제 유형 조절
+  type_adjustment:
+    enabled: true
+    weak_type_boost: 0.1     # 약한 유형 비중 증가