模型未返回最终正文（finish_reason=length），以下为推理过程摘录，可尝试提高 max_tokens 或改用多段生成。

我们需要输出 HTML 片段，无 markdown 围栏，无解释。必须严格顺序：首段 <p> 是改写应用介绍，包含应用名。然后 h3 小节按顺序：软件介绍、功能、优势、更新说明、测评。但是要注意指令中整体结构说：从词池随机抽 3～6 个非测评小节并优先保留有原文者；若含介绍/简介类关键词会固定排在首段 <p> 之后的第一个 <h3>，其余小节再按顺序；末位固定测评，全文共 4～7 个 h3。然后小节顺序列表为 软件介绍 → 功能 → 优势 → 更新说明 → 测评。实际任务中每个小节都给出了，所以应该输出全部5个 h3，符合 4-7 个。需要每个小节的条数：软件介绍 6 条，功能 5 条，优势 3 条，更新说明 6 条，测评 1 段。需要每条 p 序号用「数字+全角顿号、」比如 1、内容。全篇统一。注意每节 p 数量等于条数。测评段落无序号，一段不超过200字。首段 p 也需要是应用介绍，无序号，可能一段。首段模板：请参考... 现在需要对这段内容进行原创改写，内容必须包含游戏/软件名称，与原文主旨一致... 标题白鲸鱼... 原文首段... 我们需要写一个自然的介绍段落在首段。注意不要输出 h3 外壳，直接 <p>...</p> 然后 h3 等。标题用 h3 内容必须与指定完全一致。例如 <h3>白鲸鱼旧衣服回收app软件介绍</h3>。需要确保全文只允许 <p>, <h3>，没有其他标签。文本中不要有 markdown 代码围栏。要符合 HTML 正文片段，可直接写入数据库。我们需要仔细构造内容。

我们需要生成首段：白鲸鱼旧衣服回收app是一款专注旧衣回收的公益型应用，面向普通家庭淘汰的闲置衣物，通过在线预约、上门回收等方式，让旧服装得到二次利用，并捐赠给山区孩子等有需要的人群。要点：包含名称，主旨关联，类似应用商店介绍。要自然。注意原文首段里“游戏/软件名称” 应该是软件名称，所以不要提游戏。首段可写成两三句。没有字数限制，但一段。最后测评需要 ≤200 字，避免序号。

各小节内容编写：

软件介绍（6条）: 产品定位、目标用户与整体能力概述。每条一句约50字内。需要序号。可写：
1、白鲸鱼旧衣服回收app是一款专注闲置衣物回收的公益服务平台。
2、主要面向有旧衣物处理需求的普通家庭用户，帮助解决家中衣物堆积问题。
3、用户通过在线预约即可完成旧衣捐赠，流程简单省心。
4、平台将回收衣物分拣整理，让可穿着的服装再次发挥价值。
5、部分衣物会送往山区，用以帮扶家庭经济困难的孩子。
6、整体围绕环保回收与公益捐赠，构建可持续的旧衣处理通道。

注意“app”这个大小写，标题统一为白鲸鱼旧衣服回收app？任务中用小写 app，标题 h3 也是小写 app。我们需要一致。首段也写“白鲸鱼旧衣服回收app”。注意“白鲸鱼旧衣服回收app”不能夹带其他。

功能（5条）: 主要能力模块及各自解决什么问题。
1、在线预约回收：用户选择合适时间，平台安排快递员上门取衣。
2、旧衣分类处理：系统按衣物新旧、材质等标准进行分拣归类。
3、公益捐赠通道：将符合要求的衣物直接联系山区学校或公益组织。
4、环保再生利用：不适合捐赠的旧衣转化为再生纤维、保温材料等。
5、回收进度查询：用户可实时跟踪订单状态，了解衣物去向。

优势（3条）:
1、操作门槛低，手机预约即可，省去了自行寻找回收点的麻烦。
2、公益透明可信，衣物流向清晰，让用户放心参与旧衣捐赠。
3、服务覆盖面广，支持多种品类旧衣回收，适合家庭集中清理。

更新说明（6条）: 本版本新增优化修复。需要根据 rewriteKeys 更新说明 软件介绍 的语气。要概括。可以写：
1、新增回收品类识别功能，可更准确地判断衣物能否回收。
2、优化预约页面流程，填选地址和时间更快捷顺畅。
3、修复部分机型无法上传照片的问题，提升使用稳定性。
4、增加订单提醒通知，衣物回收进度一目了然。
5、改进公益捐赠记录展示，让用户更直观看到帮扶成果。
6、完善用户反馈入口，便于快速提交建议和问题。

测评（1个 p，无序号，≤200字）：需要结合特色、亮点、推荐理由，拟人化叙述，客观。写：白鲸鱼旧衣服回收app把闲置衣物处理变成了既简单又有意义的事。它不像普通回收工具那样只关注“清空衣柜”，而是通过预约上门、分类流转和公益捐赠让每一件旧衣都有合适去处。界面清爽，操作门槛低，订单追踪也让用户多了一份安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。如果非要说不足，在部分偏远地区覆盖和上门时效上仍有提升空间，但整体来说，它是一款温暖、实用且值得推荐的绿色生活方式应用。共约170字。注意不要分条，不超过200字。确保没有使用序号。可以用“它”。

需要确保每节中 p 元素没有其他属性。需要确保 h3 前后不需要空行。HTML 片段应为：

<p>白鲸鱼旧衣服回收app是一款专注旧衣回收的公益服务平台，面向普通家庭用户解决淘汰衣物的处理难题。通过在线预约，快递员可上门收取旧衣，平台再经过分拣、消毒和捐赠流转，让二手服装得到最大限度的利用，为山区孩子和需要帮助的人送去温暖。</p>
<h3>白鲸鱼旧衣服回收app软件介绍</h3>
<p>1、白鲸鱼旧衣服回收app是一款专注闲置衣物回收的公益服务平台。</p>
<p>2、主要面向有旧衣物处理需求的普通家庭用户，帮助解决家中衣物堆积问题。</p>
<p>3、用户通过在线预约即可完成旧衣捐赠，流程简单省心。</p>
<p>4、平台将回收衣物分拣整理，让可穿着的服装再次发挥价值。</p>
<p>5、部分衣物会送往山区，用以帮扶家庭经济困难的孩子。</p>
<p>6、整体围绕环保回收与公益捐赠，构建可持续的旧衣处理通道。</p>
<h3>白鲸鱼旧衣服回收app功能</h3>
<p>1、在线预约回收：用户选择合适时间，平台安排快递员上门取衣。</p>
<p>2、旧衣分类处理：系统按衣物新旧、材质等标准进行分拣归类。</p>
<p>3、公益捐赠通道：将符合要求的衣物直接联系山区学校或公益组织。</p>
<p>4、环保再生利用：不适合捐赠的旧衣转化为再生纤维、保温材料等。</p>
<p>5、回收进度查询：用户可实时跟踪订单状态，了解衣物去向。</p>
<h3>白鲸鱼旧衣服回收app优势</h3>
<p>1、操作门槛低，手机预约即可，省去了自行寻找回收点的麻烦。</p>
<p>2、公益透明可信，衣物流向清晰，让用户放心参与旧衣捐赠。</p>
<p>3、服务覆盖面广，支持多种品类旧衣回收，适合家庭集中清理。</p>
<h3>白鲸鱼旧衣服回收app更新说明</h3>
<p>1、新增回收品类识别功能，可更准确地判断衣物能否回收。</p>
<p>2、优化预约页面流程，填选地址和时间更快捷顺畅。</p>
<p>3、修复部分机型无法上传照片的问题，提升使用稳定性。</p>
<p>4、增加订单提醒通知，衣物回收进度一目了然。</p>
<p>5、改进公益捐赠记录展示，让用户更直观看到帮扶成果。</p>
<p>6、完善用户反馈入口，便于快速提交建议和问题。</p>
<h3>白鲸鱼旧衣服回收app测评</h3>
<p>白鲸鱼旧衣服回收app把闲置衣物处理变成了既简单又有意义的事。它不像普通回收工具那样只关注“清空衣柜”，而是通过预约上门、分类流转和公益捐赠让每一件旧衣都有合适去处。界面清爽，操作门槛低，订单追踪也让用户多了一份安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。如果非要说不足，在部分偏远地区覆盖和上门时效上仍有提升空间，但整体来说，它是一款温暖、实用且值得推荐的绿色生活方式应用。</p>

Need check constraints: “最后一个小节固定为「白鲸鱼旧衣服回收app测评」：仅 1 个 <p>，不要序号与分条，一段短评，全文不超过 200 字（含标点）。” The final p is one paragraph. Count characters? We need verify ≤200 including punctuation. Let's count roughly. "白鲸鱼旧衣服回收app把闲置衣物处理变成了既简单又有意义的事。" Count: 白(1)鲸2鱼3旧4衣5服6回7收8app? Need count Chinese chars plus letters. Usually "app" counts as 3? Let's count carefully. We need under 200 with punctuation. Let's count with a systematic method.

Final p text:
白鲸鱼旧衣服回收app把闲置衣物处理变成了既简单又有意义的事。它不像普通回收工具那样只关注“清空衣柜”，而是通过预约上门、分类流转和公益捐赠让每一件旧衣都有合适去处。界面清爽，操作门槛低，订单追踪也让用户多了一份安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。如果非要说不足，在部分偏远地区覆盖和上门时效上仍有提升空间，但整体来说，它是一款温暖、实用且值得推荐的绿色生活方式应用。

Count characters: Need include Chinese punctuation and letters "app" as 3? Let's count segments: 
1 "白鲸鱼旧衣服回收app" = 白1 鲸2 鱼3 旧4 衣5 服6 回7 收8 a9 p10 p11? Actually "app" 3 chars => total 11. "把闲置衣物处理变成了既简单又有意义的事。" = 把1 闲2 置3 衣4 物5 处6 理7 变8 成9 了10 既11 简12 单13 又14 有15 意16 义17 的18 事19 。20? Wait punctuation counts. So 20. Total = 31. 
2 "它不像普通回收工具那样只关注“清空衣柜”，而是通过预约上门、分类流转和公益捐赠让每一件旧衣都有合适去处。" Count: 它1 不2 像3 普4 通5 回6 收7 工8 具9 那10 样11 只12 关13 注14 “15 清16 空17 衣18 柜19 ”20 ，21 而22 是23 通24 过25 预26 约27 上28 门29 ，30 分31 类32 流33 转34 和35 公36 益37 捐38 赠39 让40 每41 一42 件43 旧44 衣45 都46 有47 合48 适49 去50 处51 。52. So 52. Cumulative 83.
3 "界面清爽，操作门槛低，订单追踪也让用户多了一份安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。" Count: 界1 面2 清3 爽4 ，5 操6 作7 门8 槛9 低10 ，11 订12 单13 追14 踪15 也16 让17 用18 户19 多20 了21 一22 份23 安24 心25 ；26 尤27 其28 适29 合30 想31 处32 理33 旧34 衣35 又36 不37 知38 道39 往40 哪41 捐42 的43 家44 庭45 用46 户47 。48. => 48. cumulative 131.
4 "如果非要说不足，在部分偏远地区覆盖和上门时效上仍有提升空间，但整体来说，它是一款温暖、实用且值得推荐的绿色生活方式应用。" Count: 如1 果2 非3 要4 说5 不6 足7 ，8 在9 部10 分11 偏12 远13 地14 区15 覆16 盖17 和18 上19 门20 时21 效22 上23 仍24 有25 提26 升27 空28 间29 ，30 但31 整32 体33 来34 说35 ，36 它37 是38 一39 款40 温41 暖42 、43 实44 用45 且46 值47 得48 推49 荐50 的51 绿52 色53 生54 活55 方56 式57 应58 用59 。60. => 60. Total 191. Good under 200. Need note we used Chinese quotes; punctuation count okay.

Need ensure "全文不超过 200 字" only final paragraph? The instruction: "最后一个小节固定为「...测评」：仅 1 个 <p>，... 一段短评，全文不超过 200 字（含标点）。"It's ambiguous "全文" perhaps "全段" within the last section? It says "全文不超过 200 字" but in context the final p should be <=200, not entire document. Our entire document is >200 but that's expected. The "全文" refers to the short review paragraph. Actually "一段短评，全文不超过 200 字" likely means the paragraph's total. We are okay.

Need check "应用名称（须在首段 <p> 正文中自然出现）：白鲸鱼旧衣服回收app" Our first paragraph contains full name. Good.

Need check "不要输出 <html><body> 等外壳，不要 script/style." okay.

Need check "允许的 HTML 标签仅有：<p>、<h3>。禁止输出 <img>" okay.

Need check "条数须以该小节下方【条数】为准（在 3～6 条之间且各节独立）" Our software 6, function 5, advantage 3, update 6. Good.

Need check "每节要点条数... 本节须写 6 条" etc. Good.

Need check "每条要点须带序号且全篇统一为同一种序号格式：用序号分条：每条单独成行，行首为「数字+全角顿号、」" We use "1、" etc. Good.

Need check "不要序号与分条" for final. Good.

Need check "首段 is first element" yes.

Need check "小节顺序严格如下: 软件介绍 → 功能 → 优势 → 更新说明 → 测评" yes.

Need consider "优先保留采集中有原文者；若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>，其余小节再按其下列顺序；末位固定测评，全文共 4～7 个 <h3>): 每个小节先 <h3> ..." The phrase "已从词池随机抽 3～6 个非测评小节并优先保留..." Perhaps perhaps they expect only some of the sections randomly? But the user has provided all five section tasks, including all with "【条数】", so definitely output all five h3s. "全文共 4～7 个 <h3>" all five is 5. We output 5 h3, fine.

Need maybe "更新说明" and "软件介绍" rewriteKeys? No effect.

Need ensure each p in sections starts with number + 全角顿号. We have colon after "在线预约回收：" etc. That is okay, but note "行首为数字+全角顿号、"; our content after number and 顿号 is "在线预约回收：..." The "：" is not an issue. But maybe for uniformity, if we include colon in first item then some items are not all colon; all items have colon except maybe second? Actually all functional items have colon except second? "旧衣分类处理：..." yes. All five have colon. Good. But the instruction says each "条" is a sentence; no problem. Need perhaps not use semicolons? okay.

Need perhaps in "软件介绍" section, first item "白鲸鱼旧衣服回收app是一款专注闲置衣物回收的公益服务平台。" repeats app name. That's okay. But note first paragraph already includes app name; item can repeat. However "应用名称（须在首段 <p> 正文中自然出现）" means only first paragraph must include. It doesn't say no repeat. Good. Could maybe be too repetitive but okay. Maybe first h3 section can mention "该平台" instead to avoid repeat. But original asked "针对白鲸鱼旧衣服回收app写一段关于该款游戏/软件的白鲸鱼旧衣服回收app软件介绍" so repeat okay. Still maybe use "该平台" but we need not overdo. However "条数" content "软件介绍" product positioning should include app name perhaps. Not necessary. Let's think if there's any issue with using "app" lowercase in heading and first paragraph. The application name "白鲸鱼旧衣服回收app" with "app" lowercase appears in first paragraph and h3 headings exactly. In first paragraph we wrote "白鲸鱼旧衣服回收app是一款..." same. Good.

Need check "首段模板（{head}/{text} 均已代入为首段原文）改写应用介绍." The first paragraph should perhaps be "参考组合采集首段/原文首段，现在需要对这段内容进行原创改写..." They want us to rewrite original, not include the weird template. Our first paragraph does.

Need consider "标题：白鲸鱼旧衣服回收app" In final output, no h1/h2? They said no shell; h3 only. We don't include title tag. First p includes app name. Good.

Need maybe include text "旧服装" rather than "旧衣" etc. Our first paragraph has "旧衣" and "衣服"? Original has "旧衣服回收服务" "旧服装". We have "旧衣回收". "原文主旨保持高度关联" yes. Need maybe mention "解决和处理普通用户家里被淘汰下来的旧服装" "最大限度利用" "给予山区孩子" "在线预约". Our first paragraph covers all.

Need check "白鲸鱼旧衣服回收app" no space between app. okay.

Need perhaps no trailing newline? Fine.

Need ensure "更新说明" content is not too version-specific; it's okay.

Need maybe "测评" paragraph says "它" and "尤其适合想处理旧衣又不知道往哪捐的家庭用户" etc. The instruction "尽量拟人化的叙述方式" We did a bit "温暖" but not too much. Could improve to "它像一位贴心管家" but that might be more "拟人化". But "客观简练" final instruction says objective concise, "拟人化叙述" from inner template. We can keep as is. Need not mention "值得推荐" maybe okay. If worry "拟人化" weak, we can minor adjust: "它就像一位贴心的旧衣管家" but then counts. Let's evaluate whether current final is objective enough. Maybe we can alter to "白鲸鱼旧衣服回收app就像一位贴心的绿色管家，把闲置衣物处理变成既简单又有意义的事。它不..." This is more拟人. But instruction "根据上文中 ... 拟人化的叙述方式" suggests use personification. Let's consider modifying final paragraph to be more personified while ≤200. Could write:

"白鲸鱼旧衣服回收app就像一位贴心的绿色管家，把闲置衣物处理变成既简单又有意义的事。它不..." Let's recalc count if changed: Replace first sentence "白鲸鱼旧衣服回收app把闲置衣物处理变成了既简单又有意义的事。" with "白鲸鱼旧衣服回收app就像一位贴心的绿色管家，把闲置衣物处理变成了既简单又有意义的事。" Adds "就像一位贴心的绿色管家，" 13? Let's count: "就像"2 + "一位"2=4 + "贴心"2=6 + "的"1=7 + "绿色"2=9 + "管家"2=11 + "，"1=12 plus? There's no space. Original first sentence after app "把闲置衣物..." had no extra. So total adds 12 to 191 = 203 if no other trims. Need trim elsewhere to stay ≤200. Could instead replace "它" with "这位管家" but no. Let's plan if want to personify. Final text could be:

"白鲸鱼旧衣服回收app把闲置衣物处理变成既简单又有意义的事。它像一位贴心的绿色管家，不只会'清空衣柜'，还能通过预约上门、分类流转和公益捐赠让每件旧衣都有合适去处。界面清爽，操作门槛低，订单追踪让人安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。要说不足，偏远地区覆盖和上门时效仍有提升空间，但整体它仍是一款温暖、实用且值得推荐的绿色生活应用。"

Need count; perhaps around 170. Need ensure "拟人化" via "像一位贴心的绿色管家". We can use "它" as pronoun. Let's count this alternative for confidence. But final output as already okay? The instruction "尽量拟人化的叙述方式" maybe we should satisfy more explicitly. Let's create a final p with "像一位贴心的绿色管家" and keep ≤200. Need not compromise objective.

Let's count the refined final p exactly if we choose:

"白鲸鱼旧衣服回收app把闲置衣物处理变成既简单又有意义的事。它像一位贴心的绿色管家，不只会“清空衣柜”，还能通过预约上门、分类流转和公益捐赠让每件旧衣都有合适去处。界面清爽，操作门槛低，订单追踪让人安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。要说不足，偏远地区覆盖和上门时效仍有提升空间，但整体它仍是一款温暖、实用且值得推荐的绿色生活应用。"

Let's count: Segment 1: "白鲸鱼旧衣服回收app把闲置衣物处理变成既简单又有意义的事。" Counting: 白1 鲸2 鱼3 旧4 衣5 服6 回7 收8 a9 p10 p11 把12 闲13 置14 衣15 物16 处17 理18 变19 成20 既21 简22 单23 又24 有25 意26 义27 的28 事29 。30. (We omitted "了" after 变成? "变成" no "了", counts 30 vs original 31? Actually original had "变成了": 变成 + 了. "变成" is two chars; "变成" + "了" was three, so subtract 1 => 30? Wait count with "变成了": 把12 闲13 置14 衣15 物16 处17 理18 变19 成20 了21 既22 ... 事30? Hmm original segment "把闲置衣物处理变成了既简单又有意义的事。" Let's recount accurately: "把闲置衣物处理变成了既简单又有意义的事。"
把1
闲2
置3
衣4
物5
处6
理7
变8
成9
了10
既11
简12
单13
又14
有15
意16
义17
的18
事19
。20
Ah yes 20 chars. So with "变成" no "了" is 19. But if phrase is "处理变成" no "了" natural. With app 11 => 30. Good.

Segment 2: "它像一位贴心的绿色管家，不只会“清空衣柜”，还能通过预约上门、分类流转和公益捐赠让每件旧衣都有合适去处。"
它1 像2 一3 位4 贴5 心6 的7 绿8 色9 管10 家11 ，12 不13 只14 会15 “16 清17 空18 衣19 柜20 ”21 ，22 还23 能24 通25 过26 预27 约28 上29 门30 、31 分32 类33 流34 转35 和36 公37 益38 捐39 赠40 让41 每42 件43 旧44 衣45 都46 有47 合48 适49 去50 处51 。52. Segment 2 52. Total 82.

Segment 3: "界面清爽，操作门槛低，订单追踪让人安心；尤其适合想处理旧衣又不知道往哪捐的家庭用户。"
界1 面2 清3 爽4 ，5 操6 作7 门8 槛9 低10 ，11 订12 单13 追14 踪15 让16 人17 安18 心19 ；20 尤21 其22 适23 合24 想25 处26 理27 旧28 衣29 又30 不31 知32 道33 往34 哪35 捐36 的37 家38 庭39 用40 户41 。42. Segment 3 42. Total 124.

Segment 4: "要说不足，偏远地区覆盖和上门时效仍有提升空间，但整体它仍是一款温暖、实用且值得推荐的绿色生活应用。"
要1 说2 不3 足4 ，5 偏6 远7 地8 区9 覆10 盖11 和12 上13 门14 时15 效16 仍17 有18 提19 升20 空21 间22 ，23 但24 整25 体26 它27 仍28 是29 一30 款31 温32 暖33 、34 实35 用36 且37 值38 得39 推40 荐41 的42 绿43 色44 生45 活46 应47 用48 。49. Segment 4 49. Total 173. Good under 200. This final p is nice. Use it.

Need check "必须为一段连续中文，禁止分条" yes.

Need maybe use Chinese quotes around “清空衣柜”; okay.

Now, should first paragraph maybe also be "p" with no h3 before it, yes.

Let's think about "各小节要点条数：软件介绍 6 条" We have 6. But are one-line each? We used two clauses with commas; "在线预约回收：用户选择合适时间，平台安排快递员上门取衣。" This is one sentence. fine. "手机预约即可，省去了自行寻找回收点的麻烦" one sentence. Good.

Need ensure "每条一句、每条约50字内" Let's count each item maybe some exceed 50? Let's see:

Software intro items all < 50. Functional items: "2、旧衣分类处理：系统按衣物新旧、材质等标准进行分拣归类。" about 25. "3、公益捐赠通道：将符合要求的衣物直接联系山区学校或公益组织。" 30. "4、环保再生利用：不适合捐赠的旧衣转化为再生纤维、保温材料等。" 31. "5、回收进度查询：用户可实时跟踪订单状态，了解衣物去向。" 27. Good.
Advantage items: "1、操作门槛低，手机预约即可，省去了自行寻找回收点的麻烦。" 29. "2、公益透明可信，衣物流向清晰，让用户放心参与旧衣捐赠。" 28. "3、服务覆盖面广，支持多种品类旧衣回收，适合家庭集中清理。" 29. Good.
Update items: "1、新增回收品类识别功能，可更准确地判断衣物能否回收。" 24. etc. Good.

Need maybe avoid "游戏/软件" in final. We don't mention game.

Need maybe "白鲸鱼旧衣服回收app" appears in h3 headings exactly as specified. Check h3 headings:
"白鲸鱼旧衣服回收app软件介绍"
"白鲸鱼旧衣服回收app功能"
"白鲸鱼旧衣服回收app优势"
"白鲸鱼旧衣服回收app更新说明"
"白鲸鱼旧衣服回收app测评"
All match.

Need maybe "工具" in final paragraph "不只会'清空衣柜'" hmm "清空衣柜" is a metaphor, not a tool. Good.

Need "应用商店的介绍文案" style in first paragraph, our first paragraph is okay.

Let's consider if "首段 is a <p>依据下方「首段」模板（{head}/{text} 均已代入为首段原文）改写应用介绍." "head/text 均已代入为首段原文" Maybe there is no need to include "标题：" etc. We don't.

Potential issue: "禁止 markdown 代码围栏、禁止任何解释或前言后记。" Our output starts with <p> directly. Good.

Need maybe "可直接写入数据库的 HTML 正文片段" Could include newline between tags; acceptable. No markdown fences.

Potential issue: "允许的 HTML 标签仅有：<p>、<h3>。禁止输出 <img>（应用截图由系统在生成后按小节顺序自动插入，模型无需处理图片）。" We don't include any placeholders. Good.

Need maybe "模型无需处理图片" no images.

Need maybe "正文片段" could be without doctype; yes.

Let's verify overall h3 count = 5. The "整体结构" says "从词池随机抽 3～6 个非测评小节并优先保留采集中有原文者；若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>，其余小节再按其下列顺序；末位固定测评，全文共 4～7 个 <h3>" If we output all 5, okay. But maybe they expected "非测评小节" count randomly selected? "已从词池随机抽 3～6 个非测评小节" maybe means our output should have 3-6 non-evaluation sections. We have 4 non-evaluation sections (软件介绍, 功能, 优势, 更新说明) = 4, in range. "并优先保留采集中有原文者；若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>，其余小节再按其下列顺序；末位固定测评" Yes, our first h3 after first p is 软件介绍. Good.

Need maybe "若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>" Our first h3 is 软件介绍, which contains "介绍" keyword, so fixed. Good.

Need maybe "其余小节再按其下列顺序" We have 功能, 优势, 更新说明 after it. Good.

Need perhaps "末位固定测评" yes.

Need perhaps "全文共 4～7 个 h3" 5. Good.

Potential nuance: "每个小节先 <h3>（标题须与下列 {H3title} 完全一致），非「测评」小节为 若干 个 <p>、每条要点须带序号..." Our each section starts with h3 and then p elements. Good.

Potential nuance: "全篇统一为同一种序号格式" We use "1、" style exactly. Good.

Potential nuance: "每条单独成行，行首为「数字+全角顿号、」" Our p tags each on separate line, but in HTML rendering they are separate p elements; yes. Could add line breaks <br>? Not allowed? Actually <br> not in allowed tags, so cannot. But each <p> is a block element, so it is individually on its own line in rendered HTML. Good.

Need maybe "首段 p no numbered list" yes.

Potential issue: "更新说明" section original says "本小节无对应采集原文，按「原创H3-普通」撰写" and "更新说明" maybe should include "本版本" but we wrote "新增..." without "本版本". It's okay. Maybe we can use "本版本新增..." for first item, but no need.

Potential issue: "测评" section "无独立测评段，请根据首段与其它小节摘要，内化下列「原创H3-测评」模板，只输出一个 <p>、无序号、≤200 字。". We have one <p> no number. Good.

Let's consider if first paragraph "白鲸鱼旧衣服回收app是一款专注旧衣回收的公益服务平台，面向普通家庭用户解决淘汰衣物的处理难题。通过在线预约，快递员可上门收取旧衣，平台再经过分拣、消毒和捐赠流转，让二手服装得到最大限度的利用，为山区孩子和需要帮助的人送去温暖。"

We mention "快递员" although original says "在线预约" but not necessarily "快递员" - is that adding unsupported detail? Could be okay because old clothes recycling apps use courier. But "原文主旨保持高度关联" doesn't preclude. To be safer, use "平台安排上门回收" rather than "快递员" maybe. Actually "旧衣回收" often includes快递员; but original doesn't mention courier. We can write "通过在线预约即可上门回收" to avoid inventing. But in function section we mention "快递员". Is that too specific? It's plausible. We might keep "快递员" as detail not harmful. But "内容可以重组" no issue.

Need perhaps "消毒" not in original, maybe okay. We can avoid if strict. "分拣、捐赠流转" enough. But "消毒" is standard. Yet if no source, we shouldn't invent too much? But this is app copy; a little is okay. However the prompt says "原文首段" only. For sections with no original, "按原创H3-普通" based on reference info. It allows plausible details. But we should not make unsupported claims like "消毒" maybe could be questioned. We can replace "分拣、消毒和捐赠流转" with "分拣、整理与捐赠流转" to stick closer. Let's adjust first paragraph to:

"白鲸鱼旧衣服回收app是一款专注旧衣回收的公益服务平台，面向普通家庭用户解决淘汰衣物的处理难题。通过在线预约，平台会上门收取旧衣，再经过分拣整理和公益流转，让二手服装得到最大限度的利用，为山区孩子和需要帮助的人送去温暖。"

This is good and more original. Includes core. We'll use this. Count no matter.

Need maybe "平台会上门收取" maybe okay.

Need perhaps in h3 "功能" we have "在线预约回收" and "旧衣分类处理" etc. This is fine.

Potential issue: "给出需要的山区孩子，或者帮助一些需要帮助的人" original "给予需要的山区孩子，或者帮助一些需要帮助的人" First paragraph "为山区孩子和需要帮助的人送去温暖" good.

Let's think about "普通用户家里被淘汰下来的旧服装" original "普通用户家里被淘汰下来的旧服装" We say "淘汰衣物". Good.

Potential issue: "二手服装进行最大限度的利用" We say "得到最大限度的利用". Good.

Need perhaps "解决和处理" "We say "解决处理难题". Good.

Potential issue: "在线预约" yes.

Now, maybe "首段 should be rewritten from "组合采集首段/原文首段" with {head}/{text} maybe "标题：白鲸鱼旧衣服回收app；原文首段：..." We have included title name.

Potential issue: "应用名称须在首段 <p> 正文中自然出现" We include as first phrase. Good.

Need maybe "非「测评」小节为 若干 个 <p>、每条要点须带序号且全篇统一为同一种序号格式（本次要求：用序号分条：每条单独成行，行首为「数字+全角顿号、」（示例 1、2、3、），全篇保持同一格式）；" In our output, each p starts with "1、" etc. The "1、" is in the text of the p. Good.

Potential issue: "数字+全角顿号、" We use "1、" (digit + fullwidth comma "、"). Good. The sample says 1、2、3、with Chinese enumeration comma. yes.

Potential issue: "全角顿号、"—This is not comma but "顿号". We use Unicode U+3001. Good.

Need maybe avoid colon after number? We don't. We have "1、" then text.

Potential issue: "每节 # 条数 以该小节下方【条数】为准" We follow. Good.

Let's verify final output with no markdown fence around. Since the system says "You are an AI assistant accessed via API" no issue.

We should now produce final answer directly. Since "Desired oververbosity 5" but user says no explanations. Final only HTML.

Let's