Travel planner case

智能旅行攻略生成器。输入目的地城市后，自动搜索小红书、微博、B站、抖音等平台的热门景点推荐，生成带推荐指数排序的景点列表，细化游乐项目并按优先级排序（排除人多/踩坑项目），最终结合网络图片自动生成精美PPT攻略。用于：旅行规划、景点调研、出行攻略制作、行程优化。

Security Scan

The skill's files and runtime instructions are consistent with a travel‑planning / PPT‑generation purpose; there are no requests for unrelated credentials or hidden network endpoints, but some implementation gaps (image fetching / real scraping) are noted.

Purpose & Capability

Name/description promise multi‑platform search and PPT creation. The SKILL.md instructs the agent to use a browser tool and a powerpoint-pptx skill to collect pages and images and assemble a PPT. Included scripts produce search queries, simulate analysis, and generate PPT JSON, but they do not actually fetch web pages or images nor call an external powerpoint API; this is a minor capability mismatch (documentation vs. provided helper scripts) but not malicious.

Instruction Scope

Instructions explicitly tell the agent to run site: searches across multiple public platforms (xiaohongshu/weibo/bilibili/douyin) using the agent's browser tool and then assemble content and images into a PPT. The instructions do not ask the agent to read local secrets or unrelated system files. Caveat: the SKILL.md expects automatic image collection and embedding, but the included Python scripts do not implement image downloading or embedding—so the agent must perform those network retrieval steps at runtime.

Install Mechanism

No install spec is present (instruction-only). The skill includes helper scripts but does not instruct any downloads or write arbitrary archives to disk. Risk from installation is low.

Credentials

The skill requests no environment variables, credentials, or config paths. That is proportionate for a tool that uses publicly available web searches and a PPT skill. Note: scraping some of the named platforms may require authenticated access in practice (not requested here).

Persistence & Privilege

always is false and the skill does not request permanent presence or system-wide configuration changes. The skill relies on the agent's existing browser and powerpoint capabilities to run; no elevated privileges are requested.

Assessment

This skill appears to do what it says: generate a travel PPT by collecting platform search results and assembling structured slides. Before installing or running it, consider: (1) the helper scripts only simulate search and build PPT JSON—they do not download images or perform authenticated scraping; the agent will need to perform real web retrieval and image embedding using its browser and powerpoint-pptx tools. (2) Scraping sites like 小红书/微博/抖音/B站 can be subject to login requirements and Terms of Service—ensure you have legal access and respect copyright for images. (3) Verify the agent environment provides a trusted 'browser' tool and a trustworthy powerpoint-pptx skill (to avoid unexpected data transmission). (4) Test the skill in a sandboxed account to confirm behavior, rate limits, and that no sensitive local data is sent to third parties.

Travel Planner - 智能旅行攻略生成器

本 Skill 用于自动生成完整的旅行攻略 PPT，包括多平台热门景点搜索、推荐指数排序、游乐项目优先级排序、以及自动配图生成 PPT。

使用流程

1. 接收用户输入

• 目的城市名称（必填）
• 旅行天数（可选，默认 3 天）
• 出行人数（可选）
• 预算级别（经济/舒适/豪华，可选）

2. 多平台数据搜索

使用 browser 工具搜索以下平台：

小红书:

site:xiaohongshu.com "{城市名}" "攻略" OR "推荐" OR "必去"

微博:

site:weibo.com "{城市名}" "旅游" OR "景点" "推荐"

B站:

site:bilibili.com "{城市名}" "vlog" OR "攻略" OR "游记"

抖音:

site:douyin.com "{城市名}" "打卡" OR "网红景点"

3. 数据整理与分析

从搜索结果中提取：

• 景点名称
• 推荐指数（根据出现频率、点赞数、评论热度综合评估 1-10 分）
• 标签类别（自然风光/人文历史/美食街区/现代建筑/主题乐园等）
• 游乐项目列表
• 踩坑/避雷提醒（根据负面评论提取）

4. 游乐项目优先级排序

对每个景点的游乐项目进行排序：

高优先级（推荐体验）：

• 独特性强、当地特色
• 口碑好、体验感佳
• 性价比高

中优先级（可选体验）：

• 大众化项目
• 时间充裕可体验
• 评价中等

低优先级（不建议）：

• 排队时间过长（>2小时）
• 差评集中、有明显坑点
• 性价比低、商业化过重

5. 生成 PPT

使用 powerpoint-pptx skill 生成精美 PPT：

PPT 结构：

1. 封面页 - 城市名称 + 主题标语
2. 行程概览 - 天数 + 每日主题
3. 景点推荐（按推荐指数排序）
   • 每个景点一页：名称、评分、标签、简介、精选图片
4. 游乐项目优先级表
5. 避雷指南（踩坑项目清单）
6. 实用贴士
7. 结束页

配图策略：

• 使用 browser 搜索景点图片
• 每个景点配 2-3 张高质量图片
• 使用 image attachment 或 URL 嵌入 PPT

6. 输出结果

向用户交付：

• 生成的 PPT 文件路径
• 行程亮点总结
• 重要提醒（避雷点）

关键约束

1. 数据来源多样化 - 必须搜索至少 3 个平台
2. 推荐指数客观 - 基于数据而非主观判断
3. 避雷项目明确 - 明确标注不推荐的项目及原因