caocao-chuxing-skill

Use this skill for all 曹操出行 ride-hailing requests after the user installs the skill and provides an API key from https://mcp.caocaokeji.cn/console. Trigger w...

Security Scan

The skill's code, instructions, and requirements are coherent with its stated purpose (interacting with 曹操出行 via the MCP endpoint); no evidence of unrelated or hidden behavior was found, though it will store and use an API key and perform real external actions, so you should only provide credentials to a trusted endpoint.

Purpose & Capability

Name/description match the implementation: the script calls mcp.caocaokeji.cn JSON-RPC tools for POI search, estimates, links, order create/query/cancel. Nothing in the code or SKILL.md asks for unrelated resources or credentials.

Instruction Scope

SKILL.md instructs running the included Python script to configure and call the MCP API. The instructions are scoped to POI search, estimate, link generation, and order management. Important behavioral notes in the instructions: the skill will perform real external actions (create/cancel orders) and should confirm ambiguous user intent before acting.

Install Mechanism

No install spec is provided (instruction-only plus a small included script). No downloads from external or nonstandard URLs during install; runtime network calls are to the MCP API endpoint, which is expected for this skill.

Credentials

The skill asks the user to provide an API key (via the configure command). The script writes a config file (config/mcporter.json under the current working directory) containing the base URL with the API key embedded in the query string, stored as plaintext. No other credentials or environment variables are requested.

Persistence & Privilege

always:false and no special OS privileges. The script writes its own configuration file under the current working directory but does not modify other skills or system-wide settings. The agent can call the skill autonomously (platform default).

Assessment

This skill is consistent with its stated purpose, but take these precautions before installing/using it: - Only provide your MCP/API key if you trust https://mcp.caocaokeji.cn and the organization behind it; the key is embedded in the saved baseUrl and stored in plaintext at config/mcporter.json under whatever directory the agent runs in. - Be aware that the skill performs real external actions (creating and canceling ride orders) that can result in real-world charges or commitments; the SKILL.md says to confirm with the user before executing ambiguous requests. - If you are concerned about credential leakage, run the skill in an isolated working directory, inspect or delete config/mcporter.json after use, or avoid storing long-lived keys there. - If you need higher assurance, review or host the script yourself and verify the MCP endpoint and its privacy/security policies.

曹操出行 Skill

这个 skill 用来处理 所有曹操出行打车相关需求。

安装这个 skill 后，接入者只需要做一件事：

1. 打开 https://mcp.caocaokeji.cn/console
2. 获取自己的 API Key
3. 提供给 agent 完成配置

配置完成后，就可以直接用自然语言完成：

• 搜索上车点 / 目的地
• 查看车型和价格预估
• 生成打车链接
• 直接创建订单
• 查询订单状态
• 取消订单

什么时候用

当用户出现以下意图时，直接使用这个 skill：

• “帮我叫个曹操”
• “从 A 到 B 打车”
• “查一下曹操打车多少钱”
• “生成曹操打车链接”
• “直接下单一个最便宜的车型”
• “查一下司机多久到”
• “取消这个曹操订单”
• “安装 / 配置 曹操出行 skill”

首次配置

当用户给出 API Key 时，执行：

python3 <skill_dir>/scripts/caocao_mcp.py configure --api-key "<API_KEY>"

然后建议立刻做一个简单测试：

python3 <skill_dir>/scripts/caocao_mcp.py maps_text_poi --keywords "杭州东站" --city-name "杭州"

如果返回地点列表，说明已经可以正常使用。

推荐用法

1. 高层流程

优先使用高层命令：

只估价

python3 <skill_dir>/scripts/caocao_mcp.py ride_flow --from-keywords "百度科技园" --to-keywords "北京西站" --city-name "北京" --action estimate

生成打车链接

python3 <skill_dir>/scripts/caocao_mcp.py ride_flow --from-keywords "百度科技园" --to-keywords "北京西站" --city-name "北京" --action link

直接下单（默认最便宜车型）

python3 <skill_dir>/scripts/caocao_mcp.py ride_flow --from-keywords "百度科技园" --to-keywords "百度大厦" --city-name "北京" --action create-order

指定车型下单

python3 <skill_dir>/scripts/caocao_mcp.py ride_flow --from-keywords "百度科技园" --to-keywords "北京西站" --city-name "北京" --action create-order --service-type 3

2. 低层命令

需要精确控制时，再使用：

• maps_text_poi
• trip_estimate
• trip_generate_ride_link
• trip_create_order
• trip_query_order
• trip_cancel_order

对话策略

当用户说“从 A 到 B 打车”

默认顺序：

1. 搜起点
2. 搜终点
3. 做价格预估
4. 展示车型和价格

当用户说“生成打车链接”

优先走：

• ride_flow --action link

当用户说“直接下单最便宜的车型”

优先走：

• ride_flow --action create-order
• 不指定 service-type，默认选最便宜车型

输出要求

• 地址搜索：返回前 3 个最相关 POI，带名称和经纬度
• 价格预估：清楚列车型、价格、service_type
• 链接：直接给可点击链接
• 下单成功：明确给出订单号
• 查单：优先提炼司机、车牌、距离、预计到达时间
• 取消：明确反馈是否成功

注意事项

• 下单、取消都属于真实外部动作
• 如果用户语义不明确，先确认
• 如果用户明确要求直接下单，可以直接执行