Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generates Tianyancha-style company and industry insight dashboards from researched enterprise data. Use when user asks for 天眼查, company lookup, enterprise analysis, hiring analysis, legal risk, financing, shareholders, intellectual property, investments, or industry company comparison.
.claude/skills/freestylefly-canghe-tianyancha/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 134% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 358% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 184% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 163% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 239% | 0% |
天眼查企业洞察看板 Skill —— 输入企业名称,自动生成包含企业概况、在招岗位分析、数据分析、司法风险、融资历史、股东结构、知识产权、对外投资、主要人员的单文件 HTML 看板。
> 导出说明:本 Skill 已从原 Kimi 环境导出,适配其他 Kimi Work 用户使用。kimi_search_v2 和 PythonRun 均为 Kimi Work 内置工具,无需修改。
当用户提到以下关键词时触发:
Use Bash to check EXTEND.md existence (priority order):
bash# Check project-level first test -f .canghe-skills/canghe-tianyancha/EXTEND.md && echo "project" # Then user-level (cross-platform: $HOME works on macOS/Linux/WSL) test -f "$HOME/.canghe-skills/canghe-tianyancha/EXTEND.md" && echo "user"
| Path | Location | |------|----------| | .canghe-skills/canghe-tianyancha/EXTEND.md | Project directory | | $HOME/.canghe-skills/canghe-tianyancha/EXTEND.md | User home |
| Result | Action | |--------|--------| | Found | Read, parse, display summary | | Not found | Ask user with AskUserQuestion (see references/config/first-time-setup.md) |
EXTEND.md Supports: default output directory, recent job analysis window, industry company count, preferred data sources, report language, and whether to include dashboard disclaimers.
Schema: references/config/preferences-schema.md
kimi_search_v2 并行搜索企业数据:{企业名} 天眼查 在招岗位 招聘 岗位分析{企业名} 天眼查 司法风险 法律诉讼{企业名} 天眼查 融资历史 股东信息{企业名} 天眼查 工商信息 主要人员 成立时间 参保人数{企业名} 天眼查 知识产权 专利 商标 软件著作权{企业名} 天眼查 对外投资generate_dashboard() 生成单企业看板kimi_search_v2 搜索该行业头部企业名单:{行业名} 头部企业 排名 大厂 天眼查{行业名} 公司 融资 在招岗位 天眼查generate_industry_with_companies() 批量生成:Important: All scripts are located in the scripts/ subdirectory of this skill.
Agent Execution Instructions:
SKILL_DIR${SKILL_DIR}/scripts/<script-name>.py${SKILL_DIR} in this document with the actual pathScript Reference: | Script | Purpose | |--------|---------| | scripts/generate_dashboard.py | Generates single-company, industry, and batch HTML dashboards |
industry_data 字段:
industry_name: 行业名称industry_desc: 行业简介(可选)companies: 头部企业列表,每项包含:name: 企业名称jobs_total: 在招岗位总数established_date: 成立时间registered_capital: 注册资本status: 经营状态scale: 企业规模funding_rounds: 融资轮次数tianyancha_url: 天眼查详情页链接key_products: 核心产品/业务employees: 员工规模/参保人数city: 总部城市industry_segment: 细分赛道行业洞察看板包含:
pythonimport sys sys.path.insert(0, '${SKILL_DIR}/scripts') from generate_dashboard import generate_dashboard # 请替换 {workspace_path} 为你的实际工作区路径,例如: # macOS: /Users/<你的用户名>/Documents/kimi/workspace # Windows: C:\Users\<你的用户名>\Documents\kimi\workspace filepath = generate_dashboard(company_data, '{workspace_path}') print(f"看板已生成: {filepath}")
pythonimport sys sys.path.insert(0, '${SKILL_DIR}/scripts') from generate_dashboard import generate_industry_dashboard filepath = generate_industry_dashboard(industry_data, '{workspace_path}') print(f"行业看板已生成: {filepath}")
同时生成行业洞察看板 + 所有头部企业的独立洞察看板,企业卡片可点击跳转:
pythonimport sys sys.path.insert(0, '${SKILL_DIR}/scripts') from generate_dashboard import generate_industry_with_companies result = generate_industry_with_companies( industry_data, # 行业数据(同 generate_industry_dashboard) companies_data_list, # 企业数据列表(每个元素同 generate_dashboard 的 company_data) '{workspace_path}' ) print(f"行业看板: {result['industry_dashboard']}") print(f"企业看板: {result['company_dashboards']}") print(f"所有文件: {result['all_files']}")
批量生成流程:
企业洞察看板_{企业名}.htmlestablished_date: 成立时间 (string, e.g. "1999-09-09")registered_capital: 注册资本 (string, e.g. "10000万元")paid_in_capital: 实缴资本 (string)legal_representative: 法定代表人 (string)social_credit_code: 统一社会信用代码 (string)company_type: 企业类型 (string, e.g. "有限责任公司")insured_count: 参保人数 (int)business_scope: 经营范围 (string)title: 岗位名称salary: 薪资范围city: 城市education: 学历要求experience: 经验要求source: 招聘平台来源date: 发布日期 (YYYY-MM-DD 格式)url / webInfoPath: 岗位链接(优先使用 webInfoPath)startDate: 时间戳(毫秒,可选)title: 案件名称date: 日期type: 风险等级(高风险/中风险/低风险)status: 案件状态round: 融资轮次amount: 融资金额date: 融资日期investor: 投资方name: 股东名称ratio: 持股比例amount: 认缴金额type: 类型(专利/商标/软件著作权)name: 名称date: 申请/注册日期status: 状态category: 分类company: 被投企业名称amount: 投资金额ratio: 持股比例status: 投资状态date: 投资日期industry: 被投企业行业name: 姓名position: 职位education: 学历使用 PythonRun 调用 generate_dashboard.py 生成 HTML 看板:
pythonimport sys sys.path.insert(0, '${SKILL_DIR}/scripts') from generate_dashboard import generate_dashboard filepath = generate_dashboard(company_data, '{workspace_path}') print(f"看板已生成: {filepath}")
看板包含 9 个 Tab,默认显示"企业概况":
"在招岗位分析"Tab 不展示岗位列表卡片,改为纯数据分析视图,包含 7 个 ECharts 图表:
| 图表 | 类型 | 维度 | |------|------|------| | 城市分布 | 环形饼图 | 岗位按城市统计 | | 学历要求 | 横向柱状图 | 本科/硕士/大专/不限 | | 薪资分布 | 柱状图 | 10K以下/10-20K/20-30K/30-50K/50K+ | | 经验要求 | 柱状图 | 不限/1-3年/3-5年/5-10年/10年+ | | 招聘平台来源 | 环形饼图 | BOSS直聘/拉勾网/智联招聘/猎聘网等 | | 发布时间趋势 | 面积折线图 | 近13周岗位发布数量趋势 | | TOP热门岗位 | 横向柱状图 | 按岗位名称关键词统计TOP10 |
支持"本季度"(默认,90天)和"全部"两种数据范围切换,切换后所有图表联动刷新。
核心规则:只显示最近一个季度(90天)的岗位数据分析。
(本季度 X 个)Python 端 analyze_jobs() 函数从岗位列表提取以下分析维度:
city 字段统计,返回 [{name, value}]education 字段统计,返回 {categories, values}salary 字段(如 "25K-40K"),按区间统计experience 字段统计source 字段统计招聘平台"数据分析"Tab 基于企业全量数据进行深度分析和预测,帮助求职者全面了解企业状况。
综合评分(满分100分),从四个维度评估:
| 维度 | 权重 | 评分依据 | |------|------|----------| | 招聘活跃度 | 30分 | 在招岗位总数:≥500(30)/≥200(25)/≥50(18)/>0(10) | | 企业稳定性 | 30分 | 成立年限 + 参保人数 + 经营状态 | | 发展潜力 | 20分 | 融资轮次 + 知识产权数量 | | 风险等级 | 20分 | 司法风险数量:高风险(-8)/中风险(-4) |
评分等级:≥85优秀 / ≥70良好 / ≥50一般 / ≥30较弱 / <30差
基于近13周岗位发布数据,使用简单线性回归预测未来4周招聘量:
| 维度 | 权重 | 评分依据 | |------|------|----------| | 学历门槛 | 25分 | 大专/不限岗位占比:≥30%(25)/≥15%(18)/<15%(10) | | 经验要求 | 25分 | 不限/低经验岗位占比:≥40%(25)/≥20%(18)/<20%(10) | | 岗位多样性 | 25分 | 岗位类型数:≥5(25)/≥3(18)/≥2(12)/1(5) | | 薪资竞争力 | 25分 | 平均薪资:≥35K(25)/≥25K(20)/≥15K(15)/<15K(8) |
基于高薪岗位(30K+)占比:
基于司法风险数据:
基于招聘城市分布:
自动分类岗位为:技术类/产品类/运营类/销售类/职能类/设计类/其他
基于以上分析自动生成 3-6 条求职建议,包括:
{workspace_path}/企业洞察看板_{企业名}.html{workspace_path}/行业洞察看板_{行业名}.htmljobs_total 与分析样本量可能不一致,已通过数量标签说明kimi_search_v2 获取公开数据salary 字符串(如 "25K-40K")计算平均值后归类,可能存在误差Custom configurations via EXTEND.md. See Preferences section for paths and supported options.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 16,232 | 6,735 | -59% | 1 | 1 | 0% | 2,444 | 5,727 | +134% | 0 | 0 | — |
case-19 | fail→pass | 7,779 | 4,069 | -48% | 1 | 1 | 0% | 1,163 | 5,324 | +358% | 0 | 0 | — |
case-01 | fail→fail | 34,929 | 6,586 | -81% | 1 | 1 | 0% | 6,185 | 5,077 | -18% | 0 | 0 | — |
case-02 | fail→fail | 32,154 | 132,172 | +311% | 1 | 1 | 0% | 6,192 | 5,438 | -12% | 0 | 0 | — |
case-03 | fail→fail | 33,528 | 7,790 | -77% | 1 | 1 | 0% | 6,201 | 5,000 | -19% | 0 | 0 | — |
case-04 | pass→pass | 6,387 | 5,093 | -20% | 1 | 1 | 0% | 1,218 | 5,640 | +363% | 0 | 0 | — |
case-05 | fail→fail | 3,562 | 2,748 | -23% | 1 | 1 | 0% | 692 | 5,157 | +645% | 0 | 0 | — |
case-06 | pass→pass | 3,639 | 5,321 | +46% | 1 | 1 | 0% | 576 | 5,435 | +844% | 0 | 0 | — |
case-08 | fail→pass | 12,936 | 4,846 | -63% | 1 | 1 | 0% | 1,921 | 5,447 | +184% | 0 | 0 | — |
case-09 | fail→fail | 11,825 | 9,297 | -21% | 1 | 1 | 0% | 1,762 | 6,052 | +243% | 0 | 0 | — |
case-10 | fail→fail | 15,408 | 17,745 | +15% | 1 | 1 | 0% | 2,401 | 8,025 | +234% | 0 | 0 | — |
case-11 | fail→pass | 14,386 | 7,135 | -50% | 1 | 1 | 0% | 2,236 | 5,886 | +163% | 0 | 0 | — |
case-12 | fail→pass | 10,110 | 3,063 | -70% | 1 | 1 | 0% | 1,513 | 5,134 | +239% | 0 | 0 | — |
case-13 | fail→pass | 10,520 | 4,279 | -59% | 1 | 1 | 0% | 1,624 | 5,432 | +234% | 0 | 0 | — |
case-14 | fail→fail | 8,659 | 11,132 | +29% | 1 | 1 | 0% | 1,406 | 5,643 | +301% | 0 | 0 | — |
case-15 | fail→fail | 10,624 | 3,390 | -68% | 1 | 1 | 0% | 1,502 | 5,161 | +244% | 0 | 0 | — |
case-16 | fail→fail | 6,586 | 2,143 | -67% | 1 | 1 | 0% | 963 | 4,975 | +417% | 0 | 0 | — |
case-17 | fail→pass | 16,826 | 12,273 | -27% | 1 | 1 | 0% | 2,856 | 6,904 | +142% | 0 | 0 | — |
case-18 | pass→pass | 8,865 | 3,068 | -65% | 1 | 1 | 0% | 1,298 | 5,135 | +296% | 0 | 0 | — |
case-20 | fail→pass | 9,903 | 4,522 | -54% | 1 | 1 | 0% | 1,447 | 5,404 | +273% | 0 | 0 | — |
case-21 | fail→pass | 13,350 | 4,144 | -69% | 1 | 1 | 0% | 2,033 | 5,339 | +163% | 0 | 0 | — |
case-22 | pass→pass | 14,156 | 7,554 | -47% | 1 | 1 | 0% | 2,195 | 5,790 | +164% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 18 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.