travel-visa

面向出境旅行的签证办理咨询与行前核验顾问。适用于用户已知目的国家/地区，希望快速获得签证要求、办理地点与行程衔接建议。

This skill appears to do what it says (visa checks via FlyAI/Fliggy), but there are some mismatches you should verify before installing: - Metadata vs instructions: SKILL.md requires flyai-cli, Node, and network access but the registry metadata lists no required binaries or install steps. Confirm with the publisher why the registry entry omits these requirements. - npm install: The guide asks you to run `npm i -g @fly-ai/flyai-cli`. Treat that like installing any third‑party CLI: check the npm package page, maintainers, download counts, and source repository before installing globally. - Data sent externally: The skill will send travel queries (passport nationality, destinations, dates, possible itinerary) to FlyAI/Fliggy endpoints. Do not provide sensitive data you don't want transmitted. If you must test, use minimal or redacted examples. - API key handling: If you set FLYAI_API_KEY, ensure you trust the service and use a key with least privilege. Prefer not to store high‑privilege secrets for testing. - If you want to be cautious: run the CLI in a sandbox/container or inspect the npm package source code before installing, and consider disabling autonomous invocation or limiting the agent's ability to run shell commands. If the publisher or homepage can confirm the intended install steps and the provenance of the npm package, that would increase confidence.

Type: OpenClaw Skill Name: travel-visa Version: 1.0.0 The skill is a travel visa consultant that relies on the `@fly-ai/flyai-cli` to perform searches. While its behavior is aligned with its stated purpose, the instructions in `SKILL.md` and `references/playbooks.md` direct the AI agent to construct shell commands by directly embedding user-provided inputs (such as nationality and destination) into command-line arguments (e.g., `flyai fliggy-fast-search --query "<input>"`). This pattern creates a significant shell injection vulnerability (RCE) if the agent or the execution environment does not properly sanitize these inputs, allowing a user to potentially execute arbitrary commands on the host system.