OpenClaw 存在路径遍历漏洞，Feishu 媒体临时文件命名允许在 os.tmpdir() 之外写入

该漏洞允许攻击者通过控制返回给客户端的Feishu媒体密钥值，从而影响下载文件的写入位置。由于密钥直接插入临时文件路径，攻击者可以利用路径遍历片段逃逸临时目录，将写入操作重定向到os.tmpdir()之外的任意位置。这意味着攻击者可以在OpenClaw进程的权限范围内进行任意文件写入，可能导致敏感信息泄露、恶意文件植入，甚至可能进一步提升权限，对系统造成严重损害。虽然漏洞限制在OpenClaw进程权限内，但如果OpenClaw进程具有较高的权限，则潜在影响将更大。

Organizations using OpenClaw for collaborative knowledge management, particularly those integrated with Feishu for media sharing, are at risk. Shared hosting environments where OpenClaw is deployed alongside other applications may also be vulnerable if the attacker can compromise another application and leverage it to manipulate Feishu media keys.

• nodejs: Monitor OpenClaw logs for requests containing unusual characters in the imageKey or fileKey parameters. Use grep to search for patterns like ../ or ..\ in request URLs.

grep 'imageKey=.*\.\./' /var/log/openclaw/access.log
grep 'fileKey=.*\.\./' /var/log/openclaw/access.log

• generic web: Examine access logs for requests to the media download endpoint with suspicious query parameters. Use curl to test the endpoint with crafted parameters.

curl 'https://your-openclaw-instance/download?imageKey=../../../../etc/passwd' -s

1. 2026-03-03Disclosure

   disclosure

2. 2026-03-03Patch

   patch

1. 已保留2026-01-06
2. 发布日期2026-03-03
3. 修改日期2026-03-19
4. EPSS 更新日期2026-03-25

最有效的缓解措施是立即升级到OpenClaw 2026.2.19或更高版本，该版本修复了该漏洞。如果无法立即升级，可以考虑以下临时缓解措施：首先，审查并严格控制Feishu媒体密钥的来源，确保密钥的安全性。其次，实施输入验证和清理机制，对Feishu媒体密钥进行严格的过滤和验证，防止恶意路径注入。此外，可以考虑使用Web应用防火墙（WAF）或代理服务器来拦截和过滤潜在的攻击请求。最后，定期监控系统日志，检测异常的文件写入行为。