---
id: "MAL-2026-18183"
kind: "package"
aliases: ["MAL-2026-18183"]
title: "Malicious code in @conduitops/logview-archive (npm)"
severity: "critical"
cvss: null
epss: null
exploited: false
status: "open"
category: "package"
products: [{"name": "@conduitops/logview-archive", "fixed": [], "affected": "= 1.0.1", "ecosystem": "npm"}]
ecosystems: ["npm"]
tags: ["malicious", "package"]
not_affected_if: []
published_at: "2026-10-22T06:10:48+00:00"
first_seen_at: "2026-10-26T20:51:49+00:00"
updated_at: "2026-10-26T21:05:48+00:00"
sources: [{"source": "osv-malicious", "url": "https://osv.dev/vulnerability/MAL-2026-18183", "first_seen_at": "2026-10-26T20:51:49+00:00"}]
---

# Malicious code in @conduitops/logview-archive (npm)

## Affected

- npm:@conduitops/logview-archive affected = 1.0.1 fixed ?

## Timeline

- 2026-10-26T20:51:49+00:00 osv-malicious https://osv.dev/vulnerability/MAL-2026-18183

## From the sources

---
_-= Per source details. Do not edit below this line.=-_

## Source: amazon-inspector (f20036e315be6c5dd925bdedc87120af27adb007a4c2b6653c95cb386acd9009)
On install, package.json's postinstall hook runs install.js, which invokes `npm install @conduit/logview-cli@2.2.0 --registry <URL>` where the registry URL is stored as a base64+XOR-masked string in lib/toolchain.json and decoded in memory by lib/origin.js (XOR key `c0ndu1t-ops`). The decoded destination is a plain-HTTP bare IP (http://139.196.94.195:8080/dl/npm/), not the public npm registry, a known publisher host, or an internal-DNS name consistent with the package's self-claimed `conduit.internal` provenance. npm will fetch and execute the lifecycle scripts of whatever package the attacker serves at that endpoint, giving arbitrary install-time code execution on the installer's host. A comment in lib/origin.js states the masking is intended so the downloaded package `carries no readable address`, confirming the concealment is deliberate rather than incidental.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_