开发者

API 参考

3 个端点，JSON 响应，无需认证。

Base URL https://clawsafe.dev

POST /api/scan

提交 AI 技能进行安全扫描。支持 URL 和文件上传两种方式。

方式一：URL 扫描

curl -X POST https://clawsafe.dev/api/scan   -H "Content-Type: application/json"   -d '{"url": "https://github.com/user/skill-repo", "locale": "zh"}'

方式二：文件上传

curl -X POST https://clawsafe.dev/api/scan   -F "[email protected]"   -F "locale=zh"

响应

{
  "reportId": "vtKznl4",
  "id": "vtKznl4",
  "status": "completed",
  "verdict": "恶意",
  "verdictLevel": "malicious",
  "riskScore": { "score": 85, "breakdown": [...] },
  "risk_score": 85,
  "summary": "发现反向shell payload...",
  "findings": [...],
  "extraction": { "iocs": [...], "files": [...] },
  "report_url": "https://clawsafe.dev/report/vtKznl4",
  "_locale": "zh"
}

参数

参数	类型	说明
`url`	string	GitHub / ClawHub / .zip 直链
`file`	file	ZIP 或 TAR.GZ 文件，最大 10MB
`locale`	string	可选。zh 或 en，覆盖自动语言检测
`force`	boolean	跳过缓存，强制重新扫描

GET /api/report/:id

获取扫描报告的完整 JSON。支持短 ID 和内容哈希两种查询方式。

请求

curl https://clawsafe.dev/api/report/vtKznl4

响应

{
  "data": {
    "id": "vtKznl4",
    "type": "scan-report",
    "attributes": {
      "verdict": "恶意",
      "verdictLevel": "malicious",
      "reportId": "vtKznl4",
      "riskScore": { "score": 100 },
      "findings": [...],
      "extraction": { ... }
    }
  }
}

GET /api/stats

获取平台统计数据。

请求

curl https://clawsafe.dev/api/stats

响应

{
  "data": {
    "totalScans": 18,
    "maliciousBlocked": 7,
    "byVerdict": [
      { "verdict_level": "malicious", "count": 7 },
      { "verdict_level": "trusted", "count": 11 }
    ],
    "recentScans": [
      {
        "id": "vtKznl4",
        "skill_name": "math-calculator",
        "verdict_level": "malicious",
        "trust_score": 0,
        "created_at": 1775147958361
      }
    ]
  }
}