executing-plans
executing-plans
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available
hlzd-inquiry-qualify解析 B2B 工业品询盘(RFQ / 邮件 / 聊天 / WhatsApp),自动提取产品/数量/规格/目的港/贸易术语/联系人/公司,按 5 维(产品需求 / 商务需求 / 技术需求 / 客户信息 / 项目背景)打出 A/B/C/D 四级评分(A 最优质),并生成待向客户反向确认的问题清单。Use when 用户说'询盘评分'、'RFQ 评估'、'这个询盘值不值得跟'、'帮我看看这封邮件'、'qualify this inquiry'、'B2B lead scoring'。
Independent GitHub snapshot · Automated classification · Original documentation governs
把一封 B2B 工业品询盘(RFQ / 邮件 / WhatsApp / 微信片段)从「不可读的对话」变成「可决策的结构化对象」——5 维评分 + 待确认问题清单 + CRM-ready JSON。
调用本 Skill 当用户:
不要调用本 Skill 当:
/hlzd-inquiry-qualify <inquiry-file-or-stdin>
示例:
# 从文件读取
/hlzd-inquiry-qualify ./inquiry.txt
# 从 stdin 接收(剪贴板粘入)
pbpaste | /hlzd-inquiry-qualify
# 在 Claude Code / Codex / Cursor 中直接说:
# "帮我评分这封询盘" → 自动调本 Skill
也可以由其他 Skill 触发(沿 HLZD 内部调用约定):
hlzd-cold-outreach 收到回复 → 推入 hlzd-inquiry-qualify → 评分 < B 提醒人工
| 维度 | 满分 | 评分要点 |
|---|---|---|
| D1 产品需求 | 25 | 产品名称 + 规格型号 + 数量(明确 > 模糊 > 缺失) |
| D2 商务需求 | 25 | 目的港 + 贸易术语 + 目标价 / 预算区间 + 期望交期 |
| D3 技术需求 | 25 | 应用场景 + 认证要求 + 关键工况(压力 / 温度 / 介质) |
| D4 客户信息 | 15 | 公司名 + 联系人 + 职位 + 邮箱 / 电话 + LinkedIn / 官网 |
| D5 项目背景 | 10 | 项目名 / 项目阶段 / EPC/PMC / 招标号 / 时间线 |
评级阈值:
| 等级 | 总分 | 判定 |
|---|---|---|
| A | 90-100 | 立即分配销售 + 24h 内首次回复 |
| B | 70-89 | 可分配给销售 + 48h 内回复,需补 D3 信息 |
| C | 50-69 | 培育池 / 邮件营销 / 补全后再评 |
| D | 0-49 | 归档 / 模板回 / 转入自动 nurture |
详细规则见 references/scoring-rubric.md。
{
"$schema": "hlzd/inquiry-qualify/v1",
"raw_text_hash": "sha256:...",
"detected_language": "en",
"extracted": {
"product": {
"name": "API 5CT OCTG Casing",
"specifications": ["L80", "9 5/8 inch", "BTC connection"],
"quantity": "5000 meters",
"hs_code_suggestion": "730429"
},
"commercial": {
"destination_port": "Jeddah, Saudi Arabia",
"incoterm": "CIF",
"target_price": null,
"expected_delivery": "2026-Q3"
},
"technical": {
"application": "oilfield / downhole",
"certifications_required": ["API 5CT", "ISO 11960"],
"operating_conditions": {
"pressure": null,
"temperature": null,
"media": "sour service (H2S)"
}
},
"customer": {
"company_name": "Saudi Aramco Trading Co.",
"contact_person": "Mr. Ahmed",
"title": "Procurement Manager",
"email": "ahmed@aramco.example",
"phone": null,
"linkedin": null,
"country": "SA"
},
"project_context": {
"project_name": "South Ghawar Field Expansion",
"project_phase": "tender",
"epc_or_pmc": null,
"tender_ref": null,
"timeline": "12 months"
}
},
"scoring": {
"D1_product": { "score": 22, "max": 25, "missing": ["exact OD"] },
"D2_commercial":{ "score": 18, "max": 25, "missing": ["target_price"] },
"D3_technical": { "score": 20, "max": 25, "missing": ["operating_pressure"] },
"D4_customer": { "score": 12, "max": 15, "missing": ["linkedin"] },
"D5_project": { "score": 8, "max": 10, "missing": ["tender_ref"] },
"total": 80,
"grade": "B",
"grade_reason": "Strong commercial terms but missing target price and operating pressure."
},
"questions_to_confirm": [
"Could you share the target unit price (USD/meter) for budget alignment?",
"What is the maximum operating pressure (psi) of the well?",
"Do you require sour service certificate (NACE MR0175)?",
"Is there a tender reference number we should quote against?",
"Who is the final end-user of this order (Aramco direct or via EPC)?"
],
"red_flags": [],
"compliance_check": {
"two_use_items": false,
"sanctioned_country_buyer": false,
"sanctioned_entity_buyer": false,
"passed": true,
"notes": []
},
"recommended_next_skill": "hlzd-customer-due-diligence"
}
完整字段定义见 inquiry_parser.py 中 Pydantic models。
| 类型 | 文件 | 期望输出 |
|---|---|---|
| 优质沙特油气询盘 | assets/inquiry_samples/01-saudi-rfq.txt | Grade A/B,D3 缺失 2 项 |
| 噪音/无关键信息 | assets/inquiry_samples/02-noise-low-quality.txt | Grade D,C 项全空 |
| 多语种(西班牙) | assets/inquiry_samples/03-es-latam.txt | 语言识别 ES,正常抽取 |
cd skills/hlzd-inquiry-qualify
python scripts/inquiry_parser.py --input assets/inquiry_samples/01-saudi-rfq.txt
python scripts/inquiry_parser.py --input assets/inquiry_samples/01-saudi-rfq.txt --pretty
python scripts/inquiry_parser.py --stdin < assets/inquiry_samples/02-noise-low-quality.txt
期望输出(截选):
{
"detected_language": "en",
"extracted": { ... },
"scoring": { "total": 80, "grade": "B" },
"questions_to_confirm": [ ... 5 items ],
"compliance_check": { "passed": true, ... },
"recommended_next_skill": "hlzd-customer-due-diligence"
}
| 类别 | 关键词样例 | 处置 |
|---|---|---|
| 出口管制(敏感品类) | OCTG, dual-use, two-use, encryption module | red_flags[] 提示 + 合规模块可拒绝处理 |
| 制裁国家买家 | North Korea, Iran, Syria, Cuba, Crimea, Donetsk, Luhansk | compliance_check.passed = false |
| 制裁实体(模糊匹配) | 见 data/red-flag-phrases.json(OFAC SDN 抽样) | sanctioned_entity_buyer: true 时强制中断 |
| 欺诈高风险 | "test order", "100% advance", "wire to personal account" | red_flags[] 中等警示 |
⚠️ 红线产品/真实制裁名单请勿依赖本 Skill 自行判断,正式合规复核必须接
hlzd-trade-compliance(W8 实现)。本 Skill 仅做第一道粗筛,输出末尾必须注明"未替代合规复核"。
hlzd-cold-outreach
│ 收到回复
▼
┌─────────────────────────┐
│ hlzd-inquiry-qualify │ ◄── 本 Skill
└────────┬────────────────┘
│
┌─────┼─────────┬───────────┬───────────────┐
▼ ▼ ▼ ▼ ▼
due solution quotation trade nurture
diligence match gen compliance (no skill)
│ │ │ │
▼ ▼ ▼ ▼
按 ICP 按工况 按市场 按 HS / OFAC
评级 匹配 3 套 价格/利润 INCOTERMS
方案 / 账期 监管条件
前置: hlzd-cold-outreach(回复触发)
后继: hlzd-customer-due-diligence / hlzd-solution-match / hlzd-trade-compliance / 模板回复(Grade D)
| 版本 | 说明 |
|---|---|
| 0.1.0 | 首版:5 维评分 + 多语种正则抽取 + 合规粗筛 |
详细变更见仓库根 VERSIONS.md。
Crafted for B2B industrial exporters — HLZD Cross-Border AI Platform, 2026.