目标达成 感谢每一位支持者 — 我们达成了 100% 目标！

目标: 1000 元 · 已筹: 1336 元

100%
请我们喝杯咖啡

CVE-2026-77521— MaxKB: Prompt-injectable agent can lead to command execution

一分钟漏洞结论

影响对象
1Panel-dev MaxKB
利用判断
尚无明确在野利用证据，仍需结合暴露面评估
建议动作
优先检查厂商安全公告和参考链接中的修复版本；无法立即升级时，限制受影响服务暴露并加强监测。

MaxKB is an open-source AI assistant for enterprise. Prior to version 2.10.5-lts, assistants with a tool, MCP tool, skill, or sub-application use SandboxShellBackend, which exposes an execute shell tool without excluding it and omits execute from interrupt_on,

CVSS 10.0 · Critical EPSS 1.05% · P63

影响版本矩阵 1

厂商产品 版本范围状态
1Panel-dev MaxKB < 2.10.5-lts affected
获取后续新漏洞提醒 登录后订阅

一、 漏洞 CVE-2026-77521 基础信息

漏洞信息

Shenlong is analyzing...

对漏洞内容有疑问？看看神龙的深度分析是否有帮助！
查看神龙十问 ↗

尽管我们使用了先进的大模型技术，但其输出仍可能包含不准确或过时的信息。神龙努力确保数据的准确性，但请您根据实际情况进行核实和判断。

Vulnerability Title
MaxKB: Prompt-injectable agent can lead to command execution
来源: CVE Program / CVE List V5
Vulnerability Description
MaxKB is an open-source AI assistant for enterprise. Prior to version 2.10.5-lts, assistants with a tool, MCP tool, skill, or sub-application use SandboxShellBackend, which exposes an execute shell tool without excluding it and omits execute from interrupt_on, so human approval is not required. Untrusted chat or ingested content can therefore cause command execution; source deployments with MAXKB_SANDBOX disabled run commands directly as the application user, while the official root container's string-based gosu wrapper allowed shell metacharacters to execute outside the intended sandbox. This issue is fixed in version 2.10.5-lts.
来源: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H
来源: CVE Program / CVE List V5
Vulnerability Type
OS命令中使用的特殊元素转义处理不恰当(OS命令注入)
来源: CVE Program / CVE List V5

受影响产品

厂商 产品 影响版本 CPE 订阅
1Panel-dev MaxKB < 2.10.5-lts -

二、漏洞 CVE-2026-77521 的公开POC

# POC 描述 源链接 神龙链接
AI 生成 POC 高级

未找到公开 POC。

登录以生成 AI POC

三、漏洞 CVE-2026-77521 的情报信息

请登录查看更多情报信息。

CVE-2026-77521 补丁与修复 (1)

CVE-2026-77521 其他参考 (2)

同批安全公告 · 1Panel-dev · 2026-09-21 · 共 13 条

CVE-2026-79916 9.1 CRITICAL MaxKB AWS Bedrock model credential injection leads to remote code execution
CVE-2026-77523 7.4 HIGH MaxKB: Cross-workspace model parameter form write
CVE-2026-79917 6.5 MEDIUM MaxKB: Chat share-link endpoint missing owner check: a chat token can publish another user
CVE-2026-79919 6.3 MEDIUM MaxKB function-library sandbox escape: dlopen stack-check bypass via importlib meta-path c
CVE-2026-79918 6.3 MEDIUM MaxKB: Sandbox escape via unhooked fexecve
CVE-2026-77520 5.4 MEDIUM MaxKB: Homepage ranking leaks application IDs that workflow application-nodes can use to i
CVE-2026-77517 5.4 MEDIUM MaxKB cross-knowledge IDOR lets a normal user read and modify documents and paragraphs in
CVE-2026-77516 5.4 MEDIUM MaxKB: Missing per-tool authorization in the agent and workflow tool-dispatch path
CVE-2026-77519 5.4 MEDIUM MaxKB: Expired application API keys remain usable on `/chat/api/mcp`
CVE-2026-77518 5.0 MEDIUM MaxKB: Known MCP tool IDs expose owner Tool.code and can be referenced by attacker workflo
CVE-2026-77522 4.3 MEDIUM MaxKB: Authenticated full-read SSRF via the knowledge web-document import/sync crawler (Fo
CVE-2026-77525 4.2 MEDIUM MaxKB: Management chat-record routes trust path application_id but load ChatRecord by glob

IV. Related Vulnerabilities

Same product: MaxKB
Same vendor: 1Panel-dev
Same weakness: CWE-78

V. Comments for CVE-2026-77521

暂无评论

发表评论