chore: format Phase B modules and register Phase B plan in wiki

app/question_gen/adversarial_filter.py

@@ -723,9 +723,7 @@ async def run_adversarial_rounds(
         max_steps=config.adversarial_agent_max_steps,
         model=agent.model,
     )
-    all_questions: dict[str, GeneratedQuestion] = {
-        q.question_id: q for q in initial_questions
-    }
+    all_questions: dict[str, GeneratedQuestion] = {q.question_id: q for q in initial_questions}
     pending = list(initial_questions)
     passed_now = 0
     for round_no in range(config.adversarial_max_rounds):
@@ -749,9 +747,7 @@ async def run_adversarial_rounds(
             session_id=session_id,
         )
         passed_now = len(write_final_bank(final_path, store, all_questions, cfg_fp))
-        _report_difficulty(
-            store, round_no=round_no, threshold=config.difficulty_warn_threshold
-        )
+        _report_difficulty(store, round_no=round_no, threshold=config.difficulty_warn_threshold)
         deficit = target - passed_now
         if deficit <= 0 or round_no + 1 >= config.adversarial_max_rounds:
             break

app/question_gen/run_store.py

@@ -457,16 +457,20 @@ class QuestionGenStore:
                 created_at=datetime('now')
             """,
             (
-                question_id, question_hash, stage, round, agent_prediction,
+                question_id,
+                question_hash,
+                stage,
+                round,
+                agent_prediction,
                 None if agent_correct is None else int(agent_correct),
-                verdict, pair_id, agent_config,
+                verdict,
+                pair_id,
+                agent_config,
             ),
         )
         self._conn.commit()
 
-    def completed_stages(
-        self, question_id: str, question_hash: str, agent_config: str
-    ) -> set[str]:
+    def completed_stages(self, question_id: str, question_hash: str, agent_config: str) -> set[str]:
         """返回该题在当前 hash+config 下已完成的 stage 集合（续跑用）。
 
         Parameters
@@ -501,8 +505,7 @@ class QuestionGenStore:
             当前 agent 配置指纹；保留该 config 行，其余全部删除。
         """
         self._conn.execute(
-            "DELETE FROM adversarial_verdicts "
-            "WHERE question_id=? AND agent_config!=?",
+            "DELETE FROM adversarial_verdicts WHERE question_id=? AND agent_config!=?",
             (question_id, agent_config),
         )
         self._conn.commit()
@@ -521,15 +524,12 @@ class QuestionGenStore:
             该轮 stage='cheat' 的 agent_correct 平均值；无数据时返回 0.0。
         """
         row = self._conn.execute(
-            "SELECT AVG(agent_correct) FROM adversarial_verdicts "
-            "WHERE stage='cheat' AND round=?",
+            "SELECT AVG(agent_correct) FROM adversarial_verdicts WHERE stage='cheat' AND round=?",
             (round_no,),
         ).fetchone()
         return float(row[0]) if row and row[0] is not None else 0.0
 
-    def final_passed_question_ids(
-        self, hash_by_qid: dict[str, str], agent_config: str
-    ) -> set[str]:
+    def final_passed_question_ids(self, hash_by_qid: dict[str, str], agent_config: str) -> set[str]:
         """在当前 hash+config 下通过两门的 question_id 集合（final JSON 全量重建用）。
 
         终判规则（防 stale 泄漏）：仅当该题在 **当前 question_hash + 当前
@@ -559,9 +559,7 @@ class QuestionGenStore:
             ).fetchall()
             if not rows:
                 continue
-            cheat_passed = any(
-                stage == "cheat" and verdict == "passed" for stage, verdict in rows
-            )
+            cheat_passed = any(stage == "cheat" and verdict == "passed" for stage, verdict in rows)
             no_flip = any(verdict == "filtered_no_flip" for _, verdict in rows)
             if cheat_passed and not no_flip:
                 passed.add(qid)