初始化收集对象（内部数据模型）
/webnovel-init在 Deep 采集阶段按本结构逐项收集信息；字段齐全并过「充分性闸门」后再执行webnovel.py init。本文件是字段参考，不是执行脚本。
{
"project": {
"title": "",
"genre": "",
"target_words": 0,
"target_chapters": 0,
"one_liner": "",
"core_conflict": "",
"target_reader": "",
"platform": ""
},
"protagonist": {
"name": "",
"desire": "",
"flaw": "",
"archetype": "",
"structure": "单主角"
},
"relationship": {
"heroine_config": "",
"heroine_names": [],
"heroine_role": "",
"co_protagonists": [],
"co_protagonist_roles": [],
"antagonist_tiers": {},
"antagonist_level": "",
"antagonist_mirror": ""
},
"golden_finger": {
"type": "",
"name": "",
"style": "",
"visibility": "",
"irreversible_cost": "",
"growth_rhythm": ""
},
"world": {
"scale": "",
"factions": "",
"power_system_type": "",
"social_class": "",
"resource_distribution": "",
"currency_system": "",
"currency_exchange": "",
"sect_hierarchy": "",
"cultivation_chain": "",
"cultivation_subtiers": ""
},
"constraints": {
"anti_trope": "",
"hard_constraints": [],
"core_selling_points": [],
"opening_hook": ""
}
}字段与采集步骤对应
| 字段组 | 对应 Step | 说明 |
|---|---|---|
project.* |
Step 2 | 故事核与商业定位（书名/题材/规模/一句话/核心冲突/读者/平台） |
protagonist.* |
Step 3 | 主角骨架（姓名/欲望/缺陷/原型/结构） |
relationship.* |
Step 3 | 感情线与反派分层、镜像对抗 |
golden_finger.* |
Step 4 | 金手指类型、风格、可见度、不可逆代价、成长节奏 |
world.* |
Step 5 | 世界规模、力量体系、势力、阶层、货币、境界链 |
constraints.* |
Step 6 | 反套路、硬约束、核心卖点、开篇钩子 |
必填项以 SKILL.md 的「充分性闸门」为准；可空字段允许留空或填「无」。