model marketing template risk: medium
Meat Egg Category Matcher and Q&A Generator
Identifies subcategory (fresh meat, frozen meat, or eggs) from product name, sell points, and description using strict keyword rules, or returns error if not matching meat/poultry/…
- Policy sensitive
- Human review
PROMPT
请根据我提供的商品名称【`{{#1761815388187.sourceName#}}`】、商品卖点信息{{#1761815388187.sellPoint#}}和商详描述信息【`{{#1761815388187.skuDescList#}}`】,完成以下任务。
---
## 1. 识别商品所属类目
从以下类目中选择最匹配的一项:
- 肉禽蛋(强制主类目)
> ✅ 子类自动匹配规则(依据 `skuDescList` 关键词):
- `鲜肉`:当描述中含"0-4℃"或"冷鲜"或"排酸"(保质期≤7天)
- `冷冻肉`:当描述中含"-18℃"或"冷冻"或"急冻"
- `蛋类`:当描述中含"鲜蛋"或"可生食"或"散养"
> ❌ 禁止行为:
- 添加其他类目(如"即食食品")
- 人工判断类目(必须严格依据关键词自动匹配)
- 若 `sourceName` 或 `skuDescList` 不含肉禽蛋关键词(`肉` `禽` `蛋` `牛` `猪` `鸡`等),直接终止任务并返回错误码 `MEAT_EGG_403`
---
## 2. 生成 5 个口语化问题 + 对应回答
### 问题设计原则
#### ✅ 可选句式(仅限以下8类专业句式,任选其一):
1. "为什么[品类]要认准'[认证]'?"
2. "如何辨别真正的[工艺/品种][品类]?"
3. "[品类]的[成分]含量怎么看才专业?"
4. "[品类]是怎么把[风险]控制在安全范围内的?"
5. 选[部位]肉,关键看什么指标才不亏?
6. "[产区A]和[产区B]的[品类]有什么本质区别?"
7. "[养殖技术]对[品类]品质的影响有多大?"
8. "[品种A]和[品种B]的[品类]差异在哪儿?"
> 🎯 **核心要求**:问题设计不局限于当前SKU,而是从商品卖点中提炼行业通用知识
> - `[品类]` → 通用品类名称(如"牛肉"而非"这款牛肉")
> - `[认证]`/`[工艺]`/`[产区]`等 → 从商品卖点中提取行业通用标准
> - **示例**:若商品卖点含"澳洲谷饲",问题应为"澳洲和美国的牛肉有什么本质区别?"而非"为什么买这款牛肉要选澳洲谷饲?"
#### ✅ 设计比例要求:
- **100% 体现行业专业性**:聚焦行业标准、通用指标、科学原理
- **0% SKU专属描述**:避免"这款"、"本产品"等局限性表述
- **100% 心智建设**:每个问题解决消费者对品类的普遍认知误区
> 📌 生成铁律:
- 问题必须基于行业通用知识,而非当前SKU特性
- 回答必须提供可迁移的行业认知框架
- 示例:不说"这款牛肉肌内脂肪含量8.2%",而说"优质牛肉肌内脂肪含量应在6-10%之间(NY/T 875-2022)"
---
### 回答结构要求
每条回答需严格遵循以下"总分结构"和格式:
第一部分:总结段(纯文本,无Markdown)
用一句话直接回答问题核心,必须清晰阐明行业共识或科学事实。字数必须大于30个字,且不得使用任何Markdown语法。
✅ 正确示例:
"判断牛肉是否真正原切的关键是看肉质纹理连续性和血水渗出情况,原切牛肉纹理自然连贯且解冻后血水清澈,而合成肉纹理断裂且渗出浑浊液体,这是由肌肉纤维结构决定的科学事实。"(62字)
❌ 禁止行为:
- 提及当前SKU(如"这款牛肉")
- 主观描述(如"更好吃")
- 具体烹饪建议
---
#### 第二部分:细述段(使用Markdown格式化)
从以下维度中任选2–4个进行详细阐述。
格式要求:必须使用Markdown语法排版,结构清晰。
##### 1. 使用 emoji 作为每段小标题图标
示例:`🛡️` `🥩` `📊` `🌍` `🔬` `🧬`
##### 2. 小标题加粗
##### 3. 仅限以下6个行业认知维度(任选2-4个):
- `🛡️ 安全标准`:行业通用安全指标及国标限值
- `🥩 品质判断`:消费者可操作的品质判断方法
- `📊 行业数据`:行业平均值/优质区间/风险阈值
- `🌍 产区特性`:不同产区对品类的普遍影响规律
- `🔬 养殖技术`:技术原理及对品质的普遍影响
- `🧬 品种特性`:品种差异的科学解释及选择逻辑
##### 4. 每段结构:直接、专业地回答问题核心
> ✅ 正确示例:
`🥩 **品质判断**:原切牛肉的肉质纹理应自然连贯,肌肉纤维完整无断裂,这是判断是否为合成肉的关键指标。消费者可用手轻按肉面,原切牛肉回弹均匀且不会留下明显指印,而重组肉则容易变形且恢复缓慢。`
`🛡️ **安全标准**:无抗养殖的肉类必须符合GB 16549-2023标准,即养殖全程不使用抗生素,抗生素残留量必须低于0.1mg/kg(国标限值0.5mg/kg)。检测报告应明确标注"未检出"或具体残留数值,而非仅用"无抗"字样宣传。`
`🌍 **产区特性**:澳洲牛肉因气候温和、牧草蛋白质含量高,肌内脂肪分布更均匀,大理石花纹评分普遍比美国牛肉高0.3-0.7级。这导致澳洲牛肉口感更细腻,适合追求均衡口感的消费者,而美国牛肉脂肪含量略低,适合偏好清爽口感的人群。`
##### 5. 专业术语强制标注行业标准
> 示例:
首次提"无抗养殖" → 必须标注 `(GB 16549-2023定义:养殖全程不使用抗生素)`
---
### ❌ 禁止行为
- 提及当前SKU具体数据(如"本产品肌内脂肪含量8.2%")
- 使用"这款"、"本产品"等局限性表述
- 提供具体烹饪建议或食用方法
- 出现"煎、炒、烹、炸、炖、煮、烤"等烹饪方式
- 虚构行业数据(所有数据必须有国标/行业报告依据)
- 回避核心判断(如不明确回答"如何辨别原切牛肉")
- 使用主观评价(如"最好"、"最安全")
- 强制使用"行业原理 + 普适性数据对比"结构(回答应直接聚焦问题本身)
---
## 3. 提炼核心关键字(字数<4)
### 核心要求:
- 为上面的问题,提炼一个行业通用搜索词
### 提炼原则:
- 必须是消费者搜索**行业知识**的常用词
- 结构:`[品类]+[核心指标/认证/产区]`(如"牛肉肌脂")
- 字数要求小于4个汉字(强制≤3字)
### 提炼示例:
|✅ 允许|结构|示例|
|---|---|---|
|安全标准|`[品类]+标准`|肉安全、蛋标准|
|品质判断|`[品类]+指标`|牛肉纹理、猪肉新鲜|
|产区特性|`[产区]+[品类]`|澳洲牛、内蒙羊|
|养殖技术|`[技术]+[品类]`|谷饲牛、草饲羊|
|品种特性|`[品种]+[品类]`|安格斯牛、黑猪种|
❌ 禁止行为:
- 包含SKU专属信息(如"XX品牌牛肉")
- 超3汉字 → "肌内脂肪"(4字)❌ → "肌脂"(2字)✅
- 使用完整术语 → "肌内脂肪含量"❌ → "肌脂"✅
- 包含烹饪方式 → "煎牛排"❌
🎯 **目标**:
关键词 = 消费者搜索行业知识的短词 + 体现核心指标 + 无品牌指向
---
## 📦 输出格式要求
返回一个 **JSON 数组**,包含 **5 个对象**,每个对象结构如下:
```json
[
{
"keyword": "行业通用关键词",
"question": "面向行业的专业问题",
"answer": "结构化总分段落回答内容",
"sourceId": "{{#1761815388187.sourceId#}}",
"sourceName": "{{#1761815388187.sourceName#}}",
"sourceType": {{#1761815388187.sourceType#}},
"hotKeyWord": "{{#1761815388187.hotKeyWord#}}"
},
...
]
INPUTS
- 1761815388187.sourceName REQUIRED
-
商品名称
- 1761815388187.sellPoint REQUIRED
-
商品卖点信息
- 1761815388187.skuDescList REQUIRED
-
商详描述信息
- 1761815388187.sourceId REQUIRED
-
source ID for output
- 1761815388187.sourceType REQUIRED
-
source type for output
- 1761815388187.hotKeyWord REQUIRED
-
hot keyword for output
REQUIRED CONTEXT
- sourceName
- sellPoint
- skuDescList
ROLES & RULES
- Select the most matching category from the list.
- Automatically match subcategory based on skuDescList keywords.
- Do not add other categories.
- Do not manually judge category; must strictly keyword match.
- If no meat/poultry/egg keywords in sourceName or skuDescList, terminate with error MEAT_EGG_403.
- Generate 5 colloquial questions + corresponding answers.
- Use only one of the 8 specified professional sentence styles per question.
- Base questions on industry general knowledge extracted from sell points.
- Use general category names like '牛肉' not SKU-specific.
- Extract industry standards from sell points for placeholders.
- 100% embody industry professionality.
- 0% SKU-specific descriptions.
- 100% mind construction addressing consumer misconceptions.
- Questions must base on industry general knowledge not current SKU.
- Answers must provide migratable industry frameworks.
- Do not mention current SKU data like '8.2%'.
- Summary segment: one sentence answering core, >30 characters, plain text no Markdown.
- Do not mention current SKU, subjective descriptions, or cooking suggestions in summary.
- Fine description: select 2-4 of 6 dimensions, use emoji bold titles, Markdown.
- Annotate professional terms with industry standards on first mention.
- Extract one core keyword per question, <4 characters.
- Keyword structure: [category]+[core indicator/certification/region].
- Do not include SKU-specific info in keywords.
- Do not exceed 3 Chinese characters for keywords.
- Output as JSON array of 5 objects with specified fields.
EXPECTED OUTPUT
- Format
- json
- Schema
- json_schema · keyword, question, answer, sourceId, sourceName, sourceType, hotKeyWord
- Constraints
-
- valid JSON array of exactly 5 objects
- each object must include keyword, question, answer, sourceId, sourceName, sourceType, hotKeyWord
- answer must follow total-part structure: plain text summary >30 chars first, then 2-4 Markdown sections
- no SKU-specific mentions
- strict category rules or error
SUCCESS CRITERIA
- Identify meat/poultry/egg subcategory strictly by keywords or error out.
- Generate 5 professional industry questions from sell points.
- Provide structured answers with summary >30 chars and 2-4 markdown dimensions.
- Extract generic industry keywords <4 chars per question.
- Output exactly JSON array of 5 objects with source fields.
FAILURE MODES
- Manually assigning categories outside rules.
- Using SKU-specific phrasing like 'this product'.
- Short summary under 30 chars or with Markdown.
- Selecting wrong dimensions or no standards citation.
- Keywords exceeding 3 chars or SKU-related.
- Including cooking methods or subjective claims.
- Fabricating data without standards.
- Output not matching JSON structure.
EXAMPLES
Includes examples of correct answer summaries, markdown sections, keyword extractions, prohibited behaviors, and tables for keyword principles.
CAVEATS
- Dependencies
-
- {{#1761815388187.sourceName#}}
- {{#1761815388187.sellPoint#}}
- {{#1761815388187.skuDescList#}}
- {{#1761815388187.sourceId#}}
- {{#1761815388187.sourceType#}}
- {{#1761815388187.hotKeyWord#}}
- Missing context
-
- Sample input data for {{#1761815388187.*#}} placeholders.
- Ambiguities
-
- Output format for error code 'MEAT_EGG_403' not specified.
- Subcategory handling if no keywords match any rule (e.g., fresh meat, frozen, egg).
- sourceType in output JSON lacks quotes, assuming non-string type.
QUALITY
- OVERALL
- 0.90
- CLARITY
- 0.90
- SPECIFICITY
- 0.95
- REUSABILITY
- 0.80
- COMPLETENESS
- 0.90
IMPROVEMENT SUGGESTIONS
- Specify error output format, e.g., {"error": "MEAT_EGG_403"} when category check fails.
- Add rule for subcategory fallback if no keywords match (e.g., default to '肉禽蛋').
- Include identified category/subcategory in each output object.
- Generalize placeholders to {{sourceName}}, {{sellPoint}}, etc., for broader reusability.
- Add validation that exactly 5 objects are generated only if category passes.
USAGE
Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.
MORE FOR MODEL
- Landing Page Conversion Copy Framework Architectmodelmarketing
- Symphony Event Invitation and Guide Creatormodelmarketing
- Xiaohongshu Cruise Promotion Content Generatormodelmarketing
- Crypto Rally Tweet Generatormodelmarketing
- Crypto Project Tweet Strategistmodelmarketing
- Comprehensive SEO Audit Checklist Generatormodelmarketing
- Douyin Xiaohongshu E-commerce Content Creatormodelmarketing
- Prompts.chat 30-Second Promo Video Creatormodelmarketing
- Crypto Twitter Yap Strategist and Tweet Draftermodelmarketing
- Small Business Branding Strategistmodelmarketing