Goal Reached Thanks to every supporter — we hit 100%!

Goal: 1000 CNY · Raised: 1359 CNY

100%
Buy Us a Coffee

CVE-2026-54051— Network-AI has an an OS Command Injection issue

Quick assessment

Affected
Jovancoding Network-AI
Exploitation
Public or AI PoC available; prioritize validation
Recommended action
Check the vendor advisory and references for a fixed version. If immediate upgrade is impossible, restrict exposure and increase monitoring.

Jovan Marinovic Network-AI是Jovan Marinovic个人开发者的一个 TypeScript/Node.js 多智能体编排器，可为任何 AI 智能体堆栈添加协调、防护和治理功能。 Jovan Marinovic Network-AI 5.9.1之前版本存在命令注入漏洞，该漏洞源于沙箱放行列表通配符匹配整个命令字符串，而ShellExecutor通过/bin/sh -c执行导致通配符匹配范围扩大，可能允许攻击者绕过限制执行任意命令。

CVSS 9.9 · Critical EPSS 0.67% · P51

Affected Version Matrix 1

VendorProduct Version RangeStatus
Jovancoding Network-AI < 5.9.1 affected
Get alerts for future matching vulnerabilities Log in to subscribe

I. Basic Information for CVE-2026-54051

Vulnerability Information
Have questions about the vulnerability? See if Shenlong's analysis helps!
View Shenlong Deep Dive ↗

Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.

Vulnerability Title
Network-AI has an an OS Command Injection issue
Source: CVE Program / CVE List V5
Vulnerability Description
Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.9.1, the agent sandbox gates shell commands behind an allowlist (`SandboxPolicy.isCommandAllowed`), which THREAT_MODEL.md calls the main control against a compromised agent (Adversary 3.2). The allowlist glob-matches the whole command string, but `ShellExecutor` runs that string through `/bin/sh -c`. So any wildcard allow such as `git *`, `npm *` or `node *` also matches `git status; <anything>`, and a scoped command becomes arbitrary execution. The issue is fixed in v5.9.1. `ShellExecutor` now executes via `spawn(file, args, { shell: false })` using a quote-aware parsed argv, so no shell is invoked. `SandboxPolicy.isCommandAllowed` and the new `SandboxPolicy.tokenizeCommand` reject any unquoted shell metacharacter (`; & | $ ` ` ` ( ) < > { }` newline) or unterminated quote before the allowlist glob match; quoted metacharacters are preserved as literal argument data. Users should upgrade to `network-ai@5.9.1` or later. As defense in depth, avoid broad wildcard allowlist entries such as `node *` / `npm *` which are direct code execution by design.
Source: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H
Source: 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)
Source: CVE Program / CVE List V5
Vulnerability Title
Jovan Marinovic Network-AI 命令注入漏洞
Source: CNNVD (China National Vulnerability Database)
Vulnerability Description
Jovan Marinovic Network-AI是Jovan Marinovic个人开发者的一个 TypeScript/Node.js 多智能体编排器，可为任何 AI 智能体堆栈添加协调、防护和治理功能。 Jovan Marinovic Network-AI 5.9.1之前版本存在命令注入漏洞，该漏洞源于沙箱放行列表通配符匹配整个命令字符串，而ShellExecutor通过/bin/sh -c执行导致通配符匹配范围扩大，可能允许攻击者绕过限制执行任意命令。
Source: CNNVD (China National Vulnerability Database)
CVSS Information
N/A
Source: CNNVD (China National Vulnerability Database)
Vulnerability Type
N/A
Source: CNNVD (China National Vulnerability Database)

Affected Products

Vendor Product Affected Versions CPE Subscribe
Jovancoding Network-AI < 5.9.1 -

II. Public POCs for CVE-2026-54051

# POC Description Source Link Shenlong Link
AI-Generated POC Verified env Premium
Qwen3.6-35B-A3B · 7178 chars
Pro+ exclusive includes:
Vulnerability reproduction recording (real sandbox build + trigger, exclusive)
In-depth vulnerability mechanism
Trigger conditions & impact
Full executable POC code
Exploit chain & mitigation
POC zip download
100+ AI POC generations per month

III. Intelligence Information for CVE-2026-54051

请登录查看更多情报信息。

Vendor Advisories for CVE-2026-54051 (1)

Same Patch Batch · Jovancoding · 2026-07-20 · 9 CVEs total

CVE-2026-64623 8.6 HIGH Network-AI before 5.13.4 Cryptographic Signature Verification Bypass
CVE-2026-46701 7.6 HIGH Network-AI: Unauthenticated Cross-Origin MCP Tool Invocation via Empty Default Secret
CVE-2026-64622 7.5 HIGH Network-AI 5.12.2 through 5.13.3 Missing Authorization via ApprovalInbox
CVE-2026-58484 7.1 HIGH Network-AI: Poisoned environment backup manifest allows arbitrary recursive deletion durin
CVE-2026-58481 6.5 MEDIUM Network-AI: AgentRuntime sandbox path-prefix checks allow file access outside the configur
CVE-2026-58413 6.1 MEDIUM EnvironmentManager.restore() backup ID path traversal copies arbitrary directories into en
CVE-2026-58482 5.9 MEDIUM Network-AI: ApprovalInbox HTTP server has no authentication — anyone can approve pending a
CVE-2026-58414 5.5 MEDIUM Network-AI: EnvironmentManager.backup() follows symlinked directories and copies files out

IV. Related Vulnerabilities

Same product: Network-AI
Same vendor: Jovancoding
Same weakness: CWE-78

V. Comments for CVE-2026-54051

No comments yet

Leave a comment