세션 평가·라이브코치·교수자 분석 라운드 마감 + 문서 정리 + 코드품질 리팩터

- 누적 작업트리 커밋: 회기 평가 복구·durable 저장, 라이브 코치 이력/근거, 교수자 학생분석, 음성 비언어 메타, PII 마스킹, 운영 티켓/헬스 등
- 문서: 완료 기록 docs/archive/ 냉동 보관, docs/ 단일 인덱스(docs/README.md)+통합 TODO(docs/TODO.md)로 정리
- 리팩터(행위 보존): Stage enum SSOT(taxonomy 소유·state_machine re-export), store recent/masked_turns 중복 제거, speaker_ko_label 단일 헬퍼, _list_sessions N+1 제거(state/turns 배치 + 턴평가 하이드레이션 배치)
- 검증: 백엔드 pytest 352 passed, _list_sessions E2E chromium-single-run 2 passed

apps/api/app/session_read_model.py

@@ -11,12 +11,12 @@ import re
 from collections import Counter
 from dataclasses import dataclass
 from datetime import datetime
-from typing import Literal, Optional, cast
+from typing import Any, Literal, Optional, cast
 
 from pydantic import BaseModel, Field
 
 from .config import settings
-from .services import session_metrics
+from .services import guardrail, session_metrics
 from .stage_contract import (
     ReviewPhaseKey,
     StageLabel,
@@ -31,6 +31,10 @@ WorksheetItemSpec = tuple[str, str, list[str], WorksheetSpeaker | None]
 WorksheetSectionSpec = tuple[str, str, list[WorksheetItemSpec]]
 
 LEARNER_VISIBLE_AI_ROLE = "counselor"
+MISSING_SESSION_EVALUATION_GRACE_SECONDS = 30.0
+MISSING_SESSION_EVALUATION_ERROR = (
+    "회기말 평가가 제한 시간 이후에도 저장되지 않았습니다. AI 평가 재시도가 필요합니다."
+)
 
 
 class LearnerSessionSummary(BaseModel):
@@ -816,6 +820,46 @@ def _review_summary_from_evaluation(
     return prefix + (details if details else "아래 코칭 항목은 저장된 축어록과 평가 AI 결과를 기준으로 합니다.")
 
 
+def _session_evaluation_timeout_seconds() -> float:
+    configured = float(settings.session_evaluation_timeout or settings.engine_timeout)
+    return max(1.0, configured)
+
+
+def _missing_session_evaluation_record(
+    sess: InProcSession,
+    *,
+    has_visible_turns: bool,
+    now_ts: float,
+) -> dict[str, object] | None:
+    if not sess.ended or not has_visible_turns or sess.ended_at is None:
+        return None
+    stale_after = _session_evaluation_timeout_seconds() + MISSING_SESSION_EVALUATION_GRACE_SECONDS
+    if now_ts - sess.ended_at < stale_after:
+        return None
+    return {
+        "status": "error",
+        "source": "read_model",
+        "scope": "session_end",
+        "stage": stage_label(sess.state.stage),
+        "payload": {"error": MISSING_SESSION_EVALUATION_ERROR},
+        "error": MISSING_SESSION_EVALUATION_ERROR,
+        "updated_at": iso(now_ts),
+    }
+
+
+def missing_session_evaluation_record(
+    sess: InProcSession,
+    *,
+    has_visible_turns: bool,
+    now_ts: float,
+) -> dict[str, object] | None:
+    return _missing_session_evaluation_record(
+        sess,
+        has_visible_turns=has_visible_turns,
+        now_ts=now_ts,
+    )
+
+
 def _next_line_from_evaluation(payload: dict[str, object]) -> str | None:
     alternatives = payload.get("alternative_utterances")
     if not isinstance(alternatives, list):
@@ -1025,7 +1069,26 @@ def _evaluation_payload(record: dict[str, object] | None) -> dict[str, object]:
     if not record:
         return {}
     payload = record.get("payload")
-    return payload if isinstance(payload, dict) else {}
+    if not isinstance(payload, dict):
+        return {}
+    masked = _mask_payload_text_values(payload)
+    return masked if isinstance(masked, dict) else {}
+
+
+def _mask_payload_text(value: object) -> str:
+    return guardrail.mask_pii(str(value or "")).text_masked
+
+
+def _mask_payload_text_values(value: Any) -> Any:
+    if isinstance(value, str):
+        return _mask_payload_text(value)
+    if isinstance(value, dict):
+        return {str(key): _mask_payload_text_values(child) for key, child in value.items()}
+    if isinstance(value, list):
+        return [_mask_payload_text_values(child) for child in value]
+    if isinstance(value, tuple):
+        return [_mask_payload_text_values(child) for child in value]
+    return value
 
 
 _TECHNIQUE_KIND_BY_CATEGORY = {
@@ -1091,6 +1154,17 @@ def _review_note_from_turn_eval(
     if not isinstance(ev, dict):
         return None
     quote = _review_quote_excerpt(learner_text)
+    error_text = str(ev.get("error") or "").strip()
+    if error_text:
+        return ReviewNote(
+            author="평가 AI",
+            tone="warn",
+            title="턴 평가 실패",
+            body=_review_note_body_markdown(
+                f"이 발화의 fast-loop 평가를 완료하지 못했습니다.\n\n사유: {error_text}"
+            ),
+            quote=quote,
+        )
     dev = ev.get("intent_deviation")
     if isinstance(dev, dict):
         dimension = str(dev.get("dimension") or "").strip()
@@ -1261,7 +1335,14 @@ def build_session_review(read_input: SessionReviewReadInput) -> SessionReviewRes
     if duration_seconds > 0:
         axis.append(_offset_label(duration_seconds))
 
+    now_ts = read_input.now_ts or datetime.now().timestamp()
     evaluation_record = read_input.evaluation_record
+    if evaluation_record is None and not hidden_turns:
+        evaluation_record = _missing_session_evaluation_record(
+            sess,
+            has_visible_turns=bool(visible_turns),
+            now_ts=now_ts,
+        )
     evaluation_payload = {} if hidden_turns else _evaluation_payload(evaluation_record)
     evaluation_status = (
         "" if hidden_turns else str(evaluation_record.get("status") or "") if evaluation_record else ""