SEPTEMBER 7, 2026
Live Feed
Back to database
Case File

CVE-2026-58482

MEDIUM · CVSS 5.9 EPSS 0.16% Public Exploit

Source: NVD + CISA KEV + EPSS · Published 2026-07-20 · Last synced 2026-08-19

CyberRota Analysis

AI-Generated

The `ApprovalInbox` feature in Network-AI versions 5.0.0 to 5.12.1 is vulnerable due to a lack of authentication and permissive CORS settings, allowing any party with network access to enumerate and approve pending high-risk operations without consent. This vulnerability can lead to unauthorized execution of critical actions, undermining the intended human-in-the-loop approval process. Organizations using affected versions should prioritize this issue, especially those exposing the `ApprovalInbox` to external networks, and upgrade to version 5.12.2 or later to mitigate the risk.

Public Exploit Signal

A public exploit, PoC, GitHub repository or Metasploit reference was detected for this CVE.

Note: these links are listed for security research and verification purposes only.

CVE
CVE-2026-58482
Severity
MEDIUM
CVSS
5.9
EPSS
0.16%

Original NVD Description

Network-AI, a TypeScript/Node.js multi-agent orchestrator, has a shipped, exported, documented feature called `ApprovalInbox` (`lib/approval-inbox.ts`). It is the network surface of the human-in-the-loop Approval Gate, which `ApprovalGate` uses to require explicit human approval for high-risk operations. The HTTP server it exposes has no authentication of any kind and sets `Access-Control-Allow-Origin: *` on every route, including the state-changing `POST /approvals/:id/approve` and `/deny`. As a result, in versions 5.0.0 through 5.12.1, any party who can send an HTTP request to the inbox port — a co-located process, a container/SSRF on the same host, a remote client when the operator binds a non-loopback address, or any website the operator visits in a browser (via the wildcard CORS) — can enumerate pending approvals and approve them, defeating the entire human-in-the-loop control and causing the gated high-risk action (e.g. a shell command the agent was holding for review) to execute without consent. This issue is fixed in v5.12.2. `ApprovalInbox` now accepts a `secret` option. When set, the mutating endpoints `POST /:id/approve` and `POST /:id/deny` require an `Authorization: Bearer <secret>` header, validated in constant time with `crypto.timingSafeEqual`. `startServer()` already binds to `127.0.0.1` by default; operators exposing the inbox on a network must set a secret.

Related CVEs

Other vulnerabilities affecting the same vendor(s)