Loading ToolBench
MCP server for Tencent Careers (careers.tencent.com) job search, supporting both social recruitment (社招) and campus recruitment (校招) job queries
Four tools with complete input schemas and descriptions. Naming follows verb_noun pattern (get_*, search_*). Descriptions are substantive (100-250 chars), explaining WHAT and WHEN. Parameters are well-typed with enums for constrained inputs (e.g., types array in get_job_filters, recruitType enum in search_jobs). Output schemas are not explicitly documented in source, inferred from API responses. Error handling is present but generic (catch-all in server.ts wraps exceptions as isError=true with message text). No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all tools being read-only. Pagination support present (pageIndex, pageSize) with documented limits (max 20). Tool composition is clean, each tool has one responsibility. Descriptions could better explain when to call discovery tools (get_job_filters, get_campus_filters) vs search tools.
获取腾讯校园招聘(join.qq.com)可用的筛选选项，包括招聘项目(recruitProjects，如2026校园招聘/日常实习/青云计划)、事业群(businessGroups)、职位类别(positionCategories，大类+子类两级)、工作城市(workCities，国内/海外分组)、招聘城市(recruitCities)。搜索校招职位前可先调用本工具了解可用的筛选值。
获取腾讯招聘可用的筛选选项字典，包括工作地点(WorkPlace)、事业群(BG)、职位类别(OuterType，父子两级)、招聘类型(PostAttr，如社招/校招)、国家地区(Nationality)。搜索职位前可先调用本工具了解可用的筛选值。
搜索腾讯校园招聘(join.qq.com)的在招职位，面向应届生/实习生。支持按关键词、校招项目、事业群(BG)、职位类别、工作城市筛选，筛选参数可直接传中文名称（如"深圳"、"WXG"、"软件开发类"），无需先查 code。结果分页返回，仅支持中文。
搜索腾讯招聘(careers.tencent.com)的在招职位，以社招为主。支持按关键词、城市、事业群(BG)、职位类别、招聘类型筛选，筛选参数可直接传中文名称（如"深圳"、"CSIG"、"技术研发类"），无需先查 code。结果分页返回。注意：本站点的校招类型仅含面向海外/国际生的少量岗位，官网提示"中国区的校园招聘机会请前往join.qq.com查看"，中国大陆校招/实习岗位请使用 search_campus_jobs 工具。
Output schemas not documented. Tool descriptions explain inputs but not return structure. LLMs cannot plan downstream operations or extract fields without seeing what each tool returns.
Tool annotations missing. All four tools are read-only (no state mutation), but readOnlyHint is not set. Agents cannot distinguish safe-to-retry tools from destructive ones without explicit hints.
Discovery tool descriptions lack guidance on when to call them. get_job_filters and get_campus_filters should explicitly state 'Call this first to see available filter values before calling search_jobs or search_campus_jobs.'
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | A | 82 | <=2025-11-25 | v2 |
Error messages are generic. server.ts catches all exceptions and returns isError=true with raw error text. No guidance on recovery (e.g., 'API rate limited, retry in 60s' or 'Invalid filter value, call get_job_filters to see valid options').
Parameter descriptions use Chinese examples (e.g., '深圳', 'CSIG', '技术研发类') without explaining that both Chinese names and English codes are accepted. Ambiguity about input format acceptance.