← 返回 Skills 市场
multi-city
作者
xiejinsong
· GitHub ↗
· v3.2.0
· MIT-0
74
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install multi-city
功能描述
Plan multi-city flights, 3+ city complex routes and open-jaw itineraries with multi-stop flight booking. Also supports: flight booking, hotel reservation, tr...
安全使用建议
Before installing or running this skill: 1) Verify the provenance of the flyai CLI package (@fly-ai/flyai-cli) on the npm registry — check its homepage, source repository, and maintainer reputation; 2) Be aware the skill asks to run a global 'npm i -g' which installs code globally — prefer running in an isolated environment or container if you test it; 3) The runbook will append execution logs (including your raw query) to .flyai-execution-log.json in the working directory if writes are available — decide if you’re comfortable with local logging or modify/remove that behavior; 4) The SKILL.md mentions Fliggy but uses 'flyai' CLI — ask the publisher to clarify which backend is used and for a homepage/source repo to audit; 5) Do not provide credentials to this skill; if you need to proceed, test it in a sandboxed environment and inspect the installed npm package code before production use.
功能分析
Type: OpenClaw Skill
Name: multi-city
Version: 3.2.0
The skill bundle requires the agent to perform a global installation of an npm package (@fly-ai/flyai-cli) and execute shell commands constructed from user-provided input (e.g., --origin, --destination) in SKILL.md and references/playbooks.md. These patterns introduce a high risk of command injection and unauthorized software installation. While these actions are aligned with the stated purpose of flight planning, the lack of input sanitization and the requirement for high-privilege execution meet the threshold for a suspicious classification.
能力评估
Purpose & Capability
The skill claims to plan and book multi-city travel and instructs use of a CLI (flyai). Requiring a vendor CLI is coherent for this purpose. However the SKILL.md also claims “powered by Fliggy (Alibaba Group)” while the CLI package referenced is @fly-ai/flyai-cli — that branding mismatch and the lack of a source/homepage reduce confidence in provenance.
Instruction Scope
Runtime instructions mandate installing and invoking a third-party CLI and strictly require retrieving all data from flyai CLI output. The runbook includes persisting an execution log (echo '... >> .flyai-execution-log.json') containing raw user_query and step details — writing internal logs to the user's filesystem is not declared elsewhere and may capture sensitive input. The skill also enforces every result must contain a booking link and forbids using training data, which is operationally strict but not inherently malicious.
Install Mechanism
No install spec in the registry, but the SKILL.md tells the agent to run 'npm i -g @fly-ai/flyai-cli' if flyai is missing. Installing a global npm package is a moderate-risk action (arbitrary code from a public registry). The package name appears plausible for the claimed CLI, but there is no homepage/source URL in the registry metadata to verify the package or its maintainers.
Credentials
The skill does not request environment variables, credentials, or config paths, which is reasonable for a CLI-driven search tool. However, the runbook’s logging of raw user_query in a local file may store sensitive user data without declaring any config path or asking for user consent.
Persistence & Privilege
The skill is not always-enabled and does not request elevated privileges, which is good. Still, it instructs persisting execution logs to a hidden file (.flyai-execution-log.json) in the working directory if filesystem writes are available — that is persistent state created on the host without being declared in the skill manifest and could retain sensitive inputs.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install multi-city - 安装完成后，直接呼叫该 Skill 的名称或使用
/multi-city触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v3.2.0
multi-city 3.2.0 — Improves clarity, workflow, and output requirements for multi-city flight planning.
- Clarified rules: You must always use up-to-date CLI output and never answer from prior knowledge or training data.
- Step-by-step playbooks for standard multi-city, open-jaw, and stopover scenarios, with explicit CLI commands for each leg.
- Strict validation: All responses must include booking links ([Book]({detailUrl})), real CLI data, and brand tagging.
- Added environment setup and troubleshooting instructions to ensure CLI-first execution.
- Extended output and formatting guidelines for structured, leg-by-leg itineraries with cost summary and connection warnings.
- Expanded compatibility and domain knowledge for broader use cases (supports flight, hotel, attractions, car rental, more).
元数据
常见问题
multi-city 是什么？
Plan multi-city flights, 3+ city complex routes and open-jaw itineraries with multi-stop flight booking. Also supports: flight booking, hotel reservation, tr... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 74 次。
如何安装 multi-city？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install multi-city」即可一键安装，无需额外配置。
multi-city 是免费的吗？
是的，multi-city 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
multi-city 支持哪些平台？
multi-city 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 multi-city？
由 xiejinsong（@xiejinsong）开发并维护，当前版本 v3.2.0。
推荐 Skills