Reference page — cumulative record through Wednesday, August 19, 2026 UTC. Reference pages update as the archive grows; only dated daily editions are immutable pages of record.
Jovancoding Network-AI — Network-AI: ApprovalInbox HTTP server has no authentication — anyone can approve pending agent actions
AV AC PR UI S C I A CVSS EPSS %ile KEV
N H N R U L H N 5.9 .0016 5.6 —
AFFECTED
Product Versions Fixed
Network-AI >= 5.0.0, < 5.12.2 – —
TIMELINE
Jun 30 Reserved by GitHub_M
Jul 20 EXPLOIT PUBLISHED — CVE-2026-58482 (Jovancoding Network-AI). Public exploit reference added.
Jul 20 Published (CNA: GitHub_M)
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.
Lifecycle
Complete event history — 3 events, chronological
| Date | Event | Detail |
|---|
| June 30, 2026 | Reserved | Reserved by GitHub_M |
| July 20, 2026 | EXPLOIT PUBLISHED | EXPLOIT PUBLISHED — CVE-2026-58482 (Jovancoding Network-AI). Public exploit reference added. |
| July 20, 2026 | Published | Published (CNA: GitHub_M) |
Affected
Affected products and packages — 1 row
| Vendor | Product / Package | Ecosystem | Version introduced | Fixed |
|---|
| Jovancoding | Network-AI | — | >= 5.0.0, < 5.12.2 | — |
About this page
This is a reference page, not a dated page of record. It assembles the complete lifecycle of CVE-2026-58482 from the CVE Program record, NVD enrichment, the CISA KEV catalog, EPSS, and OSV advisories. The box score's numbers (CVSS, EPSS, KEV status) are current as of Wednesday, August 19, 2026 UTC and are re-derived as the archive grows; only dated daily editions are immutable pages of record. The authoritative source for this identifier is cve.org.