> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dropcv.work/llms.txt
> Use this file to discover all available pages before exploring further.

# 快速开始（curl）

> 5 分钟跑通第一次调用。

## 前置

1. 生成 API Key（见 [鉴权](/integration/skills/authentication)）
2. 保存到本地环境变量：

```bash theme={null}
export DROPCV_API_KEY="drop_cv_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
```

## 第一次调用：搜候选人

```bash theme={null}
curl -X POST https://api.dropcv.work/api/external/v1/skills/candidate_search/invoke \
  -H "Authorization: Bearer $DROPCV_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"params": {"query": "Python 后端 5 年经验", "top_k": 5}}'
```

预期返回（截选）：

```json theme={null}
{
  "success": true,
  "skill_key": "candidate_search",
  "invocation_id": "inv_01HXXXXXXX...",
  "result": {
    "total": 12,
    "items": [
      {"candidate_id": "...", "full_name": "张三", "current_title": "高级后端", "similarity_score": 0.87, "match_summary": "..."}
    ]
  },
  "metadata": {"latency_ms": 234}
}
```

## 第二步：读完整档案

```bash theme={null}
CANDIDATE_ID="..."  # 从上一步的 items 拿
curl -X POST https://api.dropcv.work/api/external/v1/skills/candidate_profile_read/invoke \
  -H "Authorization: Bearer $DROPCV_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"params\": {\"candidate_id\": \"$CANDIDATE_ID\"}}"
```

## 下一步

* 逐个 [skill 的详细文档](/integration/skills/candidate-search)
* 在 [Claude Code / Cherry Studio 等 AI 工具里用 MCP 接入](/integration/mcp)
* 用 [`@dropcv/cli` 命令行](/integration/cli) 调用