CVE-2026-58482
Network-AI: ApprovalInbox HTTP server has no authentication — anyone can approve pending agent actions
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.
INFO
Published Date :
July 20, 2026, 5:18 p.m.
Last Modified :
Oct. 24, 2026, 8:10 p.m.
Remotely Exploit :
Yes !
Source :
[email protected]
Affected Products
The following products are affected by CVE-2026-58482
vulnerability.
Even if cvefeed.io is aware of the exact versions of the
products
that
are
affected, the information is not represented in the table below.
CVSS Scores
| Score | Version | Severity | Vector | Exploitability Score | Impact Score | Source |
|---|---|---|---|---|---|---|
| CVSS | 134c704f-9b21-4f2e-91b3-4a467353bcc0 | |||||
| CVSS 3.1 | MEDIUM | [email protected] |
Solution
- Update Network-AI to version 5.12.2 or later.
- Configure a secret for the ApprovalInbox.
- Ensure authorization headers are used for approval.
- Limit network exposure of the inbox.
References to Advisories, Solutions, and Tools
Here, you will find a curated list of external links that provide in-depth
information, practical solutions, and valuable tools related to
CVE-2026-58482.
| URL | Resource |
|---|---|
| https://github.com/Jovancoding/Network-AI/commit/a59c13a1f0ce0e8a0779a90343eef92fac5ab4c3 | Patch |
| https://github.com/Jovancoding/Network-AI/releases/tag/v5.12.2 | Patch Release Notes |
| https://github.com/Jovancoding/Network-AI/security/advisories/GHSA-mxjx-28vx-xjjj | Exploit Mitigation Patch Vendor Advisory |
| https://github.com/Jovancoding/Network-AI/security/advisories/GHSA-mxjx-28vx-xjjj | Exploit Mitigation Patch Vendor Advisory |
CWE - Common Weakness Enumeration
While CVE identifies
specific instances of vulnerabilities, CWE categorizes the common flaws or
weaknesses that can lead to vulnerabilities. CVE-2026-58482 is
associated with the following CWEs:
Common Attack Pattern Enumeration and Classification (CAPEC)
Common Attack Pattern Enumeration and Classification
(CAPEC)
stores attack patterns, which are descriptions of the common attributes and
approaches employed by adversaries to exploit the CVE-2026-58482
weaknesses.
We scan GitHub repositories to detect new proof-of-concept exploits. Following list is a collection of public exploits and proof-of-concepts, which have been published on GitHub (sorted by the most recently updated).
Results are limited to the first 15 repositories due to potential performance issues.
The following list is the news that have been mention
CVE-2026-58482 vulnerability anywhere in the article.
The following table lists the changes that have been made to the
CVE-2026-58482 vulnerability over time.
Vulnerability history details can be useful for understanding the evolution of a vulnerability, and for identifying the most recent changes that may impact the vulnerability's severity, exploitability, or other characteristics.
-
CVE Translated by [email protected]
Oct. 24, 2026
Action Type Old Value New Value Added Translation Title: Network-AI de Jovancoding, Description: Network-AI, un orquestador multiagente de TypeScript/Node.js, tiene una característica distribuida, exportada y documentada llamada ’ApprovalInbox’ (’lib/approval-inbox.ts’). Es la superficie de red de la Puerta de Aprobación con intervención humana, que ’ApprovalGate’ utiliza para requerir aprobación humana explícita para operaciones de alto riesgo. El servidor HTTP que expone no tiene autenticación de ningún tipo y establece ’Access-Control-Allow-Origin: *’ en cada ruta, incluyendo las que cambian de estado ’POST /approvals/:id/approve’ y ’/deny’. Como resultado, en las versiones 5.0.0 a 5.12.1, cualquier parte que pueda enviar una solicitud HTTP al puerto de la bandeja de entrada - un proceso colocalizado, un contenedor/SSRF en el mismo host, un cliente remoto cuando el operador vincula una dirección no-loopback, o cualquier sitio web que el operador visite en un navegador (a través del CORS comodín) - puede enumerar las aprobaciones pendientes y aprobarlas, anulando todo el control con intervención humana y haciendo que la acción de alto riesgo protegida (p. ej., un comando de shell que el agente estaba reteniendo para revisión) se ejecute sin consentimiento. Este problema está solucionado en la v5.12.2. ’ApprovalInbox’ ahora acepta una opción ’secret’. Cuando se establece, los endpoints mutantes ’POST /:id/approve’ y ’POST /:id/deny’ requieren un encabezado ’Authorization: Bearer [secret]’, validado en tiempo constante con ’crypto.timingSafeEqual’. ’startServer()’ ya se vincula a ’127.0.0.1’ por defecto; los operadores que exponen la bandeja de entrada en una red deben establecer un secreto. -
Initial Analysis by [email protected]
Aug. 13, 2026
Action Type Old Value New Value Added CPE Configuration OR *cpe:2.3:a:network-ai:network-ai:*:*:*:*:*:*:*:* versions from (including) 5.0.0 up to (excluding) 5.12.2 Added Reference Type GitHub, Inc.: https://github.com/Jovancoding/Network-AI/commit/a59c13a1f0ce0e8a0779a90343eef92fac5ab4c3 Types: Patch Added Reference Type GitHub, Inc.: https://github.com/Jovancoding/Network-AI/releases/tag/v5.12.2 Types: Patch, Release Notes Added Reference Type GitHub, Inc.: https://github.com/Jovancoding/Network-AI/security/advisories/GHSA-mxjx-28vx-xjjj Types: Exploit, Mitigation, Patch, Vendor Advisory Added Reference Type CISA-ADP: https://github.com/Jovancoding/Network-AI/security/advisories/GHSA-mxjx-28vx-xjjj Types: Exploit, Mitigation, Patch, Vendor Advisory -
CVE Modified by 134c704f-9b21-4f2e-91b3-4a467353bcc0
Jul. 20, 2026
Action Type Old Value New Value Added Reference https://github.com/Jovancoding/Network-AI/security/advisories/GHSA-mxjx-28vx-xjjj Added SSVC {'id': 'CVE-2026-58482', 'role': 'CISA Coordinator', 'options': [{'exploitation': 'poc'}, {'automatable': 'no'}, {'technicalImpact': 'partial'}], 'version': '2.0.3', 'timestamp': '2026-07-20T17:33:19.793160Z'} -
New CVE Received by [email protected]
Jul. 20, 2026
Action Type Old Value New Value Added Affected [{'vendor': 'Jovancoding', 'product': 'Network-AI', 'versions': [{'status': 'affected', 'version': '>= 5.0.0, < 5.12.2'}]}] Added 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. Added CVSS V3.1 AV:N/AC:H/PR:N/UI:R/S:U/C:L/I:H/A:N Added CWE CWE-352 Added CWE CWE-862 Added Reference https://github.com/Jovancoding/Network-AI/commit/a59c13a1f0ce0e8a0779a90343eef92fac5ab4c3 Added Reference https://github.com/Jovancoding/Network-AI/releases/tag/v5.12.2 Added Reference https://github.com/Jovancoding/Network-AI/security/advisories/GHSA-mxjx-28vx-xjjj