MCP server providing access to Customs big data APIs for enterprise trade information queries including export trends, product trends, recruitment profiles, overseas certifications, and fuzzy search capabilities.
This server exposes 6 customs data query tools via FastMCP with STDIO transport. While tool names follow a naming convention (customs_bigdata_*) and descriptions are present, the implementation has significant quality gaps: (1) input schemas are inferred from Python type hints and docstrings, not explicitly visible as JSON Schema objects; (2) parameter descriptions lack critical constraints (enums, ranges, validation rules); (3) no explicit error handling or recovery guidance; (4) output schemas are documented only in docstring comments, not in machine-readable form; (5) enum values like keywordType are described in prose but not enforced as constrained inputs; (6) parameter relationships (e.g., matchKeyword + keywordType dependency) are undocumented. All tools are READ_ONLY and functionally similar (query customs data), which is appropriate, but the execution lacks production-grade tooling patterns.
该接口的功能是根据提供的企业标识信息(如企业名称、注册号等)查询该企业外贸订单的地理分布情况,输出包括订单地区列表、订单金额、订单数量、涉及的国家或地区、净重,以及分布地区的总数。此接口可用于企业管理系统中,帮助企业分析其国际市场的分布状况,评估各地区的订单贡献度,从而制定更加有效的市场拓展策略。
该接口用于查询并分析企业外贸商品的发展趋势,提供企业在不同年份的出口商品订单量及海关HS编码分布信息。其功能包括通过输入企业的基本标识信息(如企业名称、注册号等),输出该企业历年出口商品的订单量变化情况及具体商品类别。
该接口的功能是根据输入的企业标识信息(如企业名称或统一社会信用代码)和主体类型枚举,获取企业在过去三个月内的外贸相关岗位的招聘情况。输出信息包括招聘岗位数量、招聘城市及数量、招聘渠道和外贸相关的招聘渠道名称等。此接口可能在以下场景中使用:企业想要了解竞争对手的外贸招聘动态,以帮助其优化招聘策略;政府机构分析某地区外贸企业的用工需求变化;招聘平台分析外贸行业的岗位需求趋势,以促进精准服务。
该接口的功能是根据企业的身份标识信息(如企业名称、注册号、统一社会信用代码等)查询该企业的外贸出口趋势,包括每年的出口金额和订单数量等数据。此接口适用于分析企业在国际市场中的表现,帮助企业管理者或研究人员了解企业的出口能力以及市场份额的变化趋势。具体场景包括企业在进行市场分析时需要了解自身或竞争对手在外贸方面的增长潜力、银行或投资机构在金融评估时用于判断企业的国际贸易风险和收益、以及政府或行业协会在统计和研究企业整体外贸出口趋势时使用。
该接口的功能是根据提供的企业名称、人名、品牌、产品、岗位等关键词模糊查询相关企业列表。返回匹配的企业列表及其详细信息,用于查找和识别特定的企业信息。
Input schemas not explicitly defined as JSON Schema objects. Parameter constraints (enums, ranges) are documented only in prose (docstrings), not enforced via schema. LLMs cannot validate constrained inputs (e.g., keywordType must be one of: name, nameId, regNumber, socialCreditCode) without explicit enum definitions.
Output schemas documented only in docstring comments (plain text), not in machine-readable JSON Schema format. Agents cannot programmatically validate or route response fields. Documentation is human-readable but not machine-parseable.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 49 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 36 | - | v1 |
该接口的功能是根据企业的名称、注册编号、统一社会信用代码或企业ID查询企业的海外认证信息,包括认证类别、证书编号及相关产品信息。该接口可用于企业在寻求国际贸易机会时评估潜在合作伙伴的资质,帮助中介公司或行业分析者收集市场数据,或者企业在申请招投标过程中提交其资质证明文档。通过提供认证信息,还可方便监管机构和行业组织核实企业的合规性和认证情况。
No explicit error handling or recovery guidance in tool code. call_api() function returns generic strings ('接询失败', '接口调用失败') with no actionable next steps. LLMs cannot self-correct or retry intelligently.
Parameter relationships undocumented. All tools accept matchKeyword + keywordType as co-dependent parameters (keywordType determines how matchKeyword is interpreted), but the docstrings do not explain this relationship or make keywordType truly optional with smart defaults.
No input validation or sanitization visible in tool code. Credentials (INTEGRATOR_ID, SECRET_ID, SECRET_KEY) are loaded from environment variables correctly, but no explicit checks for injection attacks or malformed API payloads before forwarding to call_api().
Pagination parameters (pageIndex, pageSize) present in only 2 of 6 tools (export_order_regions, fuzzy_search). Other tools may return large result lists without pagination support, risking context window exhaustion.