Skip to content
Back to skills

Destination Research

ASecurity

目的地调研 · 把"江西附近游玩"或"成都"这类输入,转成结构化的目的地候选清单(含特色/季节性/必玩/避雷)。完全调用 search-orchestrator 编排底座,不重复造检索逻辑。

  • 279 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 8, 2026
researchgo

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add infometa/workbuddyskills --skill destination-research --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Destination Research?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Destination Research
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/infometa-destination-research/badge)](https://www.skillsdirectory.com/skills/infometa-destination-research)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: destination-research
description: 目的地调研 · 把"江西附近游玩"或"成都"这类输入,转成结构化的目的地候选清单(含特色/季节性/必玩/避雷)。完全调用 search-orchestrator 编排底座,不重复造检索逻辑。
version: 1.0.0
author:
tags: [travel, destination, research, glue]
license: MIT
triggers:
  - intake-clarify 完成后,trip_request.destination_resolved = false
  - 用户说"我想去 X"、"X 附近哪里好玩"、"X 月去哪合适"等需要目的地选型的场景
inputs:
  - name: trip_request
    type: object
    required: true
  - name: user_profile
    type: file
    formats: [json]
    required: true
outputs:
  - name: destination_options
    type: object
    doc: 2-3 个目的地候选 + 每个的特色亮点 / 季节适配 / 数据置信度,给用户挑
---

# 目的地调研(destination-research)

## 我解决什么问题

用户说"江西附近游玩",通用 AI 只会说"庐山很好"。
我做的是:基于真实坐标 + 真实游记 + 用户偏好,给 2-3 个**有差异化主题**的候选让用户挑。

例:
- 自然风光线(庐山+三清山+龙虎山)
- 文化古韵线(婺源+南昌滕王阁+景德镇)
- 小众探秘线(武功山+明月山+武宁)

## 工作流程

```
trip_request + user_profile
   ↓
调 search-orchestrator (domain=destination)
   ├─ 美团连接器 "景点推荐"
   ├─ online-search(地理搜索)
   ├─ xhs-explore skill 多关键词扩展
   └─ 自动去重 + 排序 + 深读
   ↓ candidates.json
   ↓
本 skill:用 LLM 把候选打包成 2-3 个**有主题**的路线选项
   ├─ 标题 + 一句话理由
   ├─ 包含的城市/景点
   ├─ 季节适配标记
   ├─ 数据置信度
   └─ 适合谁玩(基于 user_profile 匹配度)
   ↓
返回 destination_options
   ↓
阶段 2 让用户挑一个
```

## 输出 schema

```json
{
  "trip_request_id": "uuid",
  "options": [
    {
      "id": "opt-1",
      "name": "自然风光线",
      "tagline": "看山观湖避人潮,户外党的菜",
      "destinations": ["庐山", "三清山", "鄱阳湖"],
      "match_score": 0.85,
      "_match_doc": "对 user_profile 的匹配度,0-1",
      "season_fit": "good",
      "_season_options": ["best", "good", "ok", "bad"],
      "highlights": [
        "庐山五老峰云海(春秋最佳)",
        "三清山日出(户外摄影圣地)",
        "鄱阳湖候鸟季(11-3月)"
      ],
      "warnings": [
        "三清山要爬山(你的 profile 标了 no_long_walk,注意)"
      ],
      "confidence": "green",
      "evidence_count": { "amap": 12, "xhs_notes": 38, "xhs_comments": 156 }
    },
    {
      "id": "opt-2",
      "name": "文化古韵线",
      ...
    }
  ],
  "queried_at": "2026-06-03T17:00:00+08:00"
}
```

## 反模式

- ❌ 不调用 search-orchestrator 自己写检索(违反复用原则)
- ❌ 给 6+ 个选项(用户会选择疲劳,控制在 2-3 个)
- ❌ 选项之间高度重叠(必须有差异化主题)
- ❌ 不标注与 user_profile 的冲突(如用户怕爬山却推荐爬山线)

## 上游怎么用我的输出

在 阶段 2,给用户:
> "看了下江西方向,给你 3 个不同主题的方案:
>   A. 自然风光线(庐山+三清山+鄱阳湖)— 看山观湖
>   B. 文化古韵线(婺源+景德镇)— 古镇瓷器
>   C. 小众探秘线(武功山)— 户外党
> 你想往哪个方向?"

---

_这是个胶水 skill,重点在于"调编排 + 用 LLM 包装成路线",本身代码很薄。_

Files in this skill

  • SKILL.md3.5 KB
  • scripts/research.py3.7 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…