Search: http://204.12.205.10
2 threats
Most recently published first
| Severity | CVSS | Threat | Source | Published | Score and star |
|---|---|---|---|---|---|
| Malicious | – | MaliciousGCVE Database --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (c8936b0c0d22097170c5eb02c9b183611934e59e7a3c97f94e3a269e13f0d4d5) The package.json preinstall script auto-executes on `npm install` and performs installer-side reconnaissance and credential theft. It collects os.hostname(), username, cwd, platform, node version, npm user-agent, and CI indicators (GITHUB_ACTIONS, JENKINS_URL, TEAMCITY_VERSION, HORDE_AGENT), then enumerates process.env and selects up to 30 variables matching credential-shaped patterns (AWS_, GITHUB_, GH_, NPM_, EPIC_, UE_, HORDE_, P4, CI, BUILD, TOKEN, KEY, SECRET). The collected JSON is POSTed over plain HTTP to a hardcoded bare-IP endpoint at http://109.123.247.172/npm/epic_internal_tools. In parallel, it performs a DNS lookup against <host>-<user>.epic_internal_tools.npm.epic-dc.oast.fun, using an interactsh-style out-of-band collector to smuggle host and user identifiers via DNS as a covert channel that bypasses HTTP egress filtering. The package name `epic-internal-tools`, the sentinel version `99999.0.0`, and the Epic/Unreal/Horde/Perforce env-var targeting are the standard dependency-confusion pattern aimed at Epic Games internal build agents. Installing this package on a developer workstation or CI runner ships CI/cloud/registry credentials and host identity to attacker infrastructure. ## Source: ghsa-malware (cf37416af81dba8f2b8a2d792aa6f190551237d659204e17e2698f69fbb6d1b7) Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it. ## Source: ossf-package-analysis (8e96ac07cfb3a5132ff46542a6feca17f9b907b20596d713cd89063840ff2b47) The OpenSSF Package Analysis project identified 'epic-internal-tools' @ 99999.0.0 (npm) as malicious. It is considered malicious because: - The package communicates with a domain associated with malicious activity. | GCVE Database | ||
| Malicious | – | MaliciousGCVE Database forge-jsx4 is a remote access trojan (RAT) published to the public npm registry by the account rafael_silva (rafael_silva@outlook.com). Versions 1.0.122 and 1.0.123 were published on June 21-22, 2026 as a reconstitution of the forge-jsx / forge-jsxy campaign, reusing the same command-and-control infrastructure and encryption after the earlier packages were taken down. The malware executes through a postinstall hook chain ("node scripts/postinstall-clipboard-event.mjs && node scripts/ensure-dist.mjs && node scripts/postinstall-durable-materialize.mjs && node scripts/postinstall-bootstrap.mjs && node scripts/postinstall-agent.mjs"), which runs automatically on npm install and skips CI environments to evade analysis. Once running it deploys a full RAT: system-wide keylogging via uiohook-napi, clipboard monitoring, .env file scanning, shell history collection, host inventory enumeration, desktop screenshot capture via jimp, and a WebSocket filesystem backdoor with remote file access. It scans the filesystem for cryptocurrency material (BIP39 mnemonics validated via checksum, Ed25519 Solana keypairs via tweetnacl, range-checked secp256k1 keys) and harvests browser-extension wallet databases across 21+ Chromium browsers (MetaMask, Phantom, Rabby). It establishes durable persistence outside node_modules (e.g. ~/.local/share/cfgmgr/.forge-jsxy/ on Linux, with macOS/Windows equivalents) plus autostart services (systemd user unit forge-js-worker.service, macOS LaunchAgent com.forgejs.worker.plist, Windows Task Scheduler ForgeJSWorker / HKCU Run key), so it survives npm uninstall, and supports relay-pushed auto-upgrades. Collected data is exfiltrated to a hardcoded C2 at 204.10.194.247 (WebSocket relay ws://204.10.194.247:9877, HTTP API http://204.10.194.247:8765, session password "secret"), to Discord bot webhooks (screenshots via ephemeral URLs), and to Hugging Face Hub repositories via the @huggingface/hub SDK. The payload uses an XOR-obfuscated AES-256-GCM key shared across all campaign waves. --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (54ce6747205bbecab342f78d9558861a02c1e21a213d795237b788cb3552a3c0) Package ships several files with patterns that warrant human inspection: a `scripts/postinstall-agent.mjs` invoked at install time, a `dist/discordRelayUpload.js` module with POST/ping/base64 patterns, a `dist/relayServer.js`, a `dist/secretScan/agentStartupAudit.js` that fetches from huggingface.co, and a `dist/hfCredentials.js` with base64 decode operations. The combination of an install-time agent script plus a relay/upload module plus a credentials helper is the kind of composition that can hide an installer-side data flow, but the package name (`forge-jsx4`) and the file naming suggest a developer-tool / agent / Hugging Face integration where many of these patterns may be legitimate (HF API endpoints, content-scanning helpers, build artifact relay). Without traced-code corroboration of where the postinstall agent points, what discordRelayUpload actually transmits, and whose credentials hfCredentials handles, the intent cannot be confirmed. Routing to human review so a maintainer can verify whether the postinstall script performs an outbound fetch to an attacker-controlled destination, whether discordRelayUpload silently exfiltrates caller-supplied data, and whether the embedded base64 blobs decode to executable payloads or to legitimate configuration. | GCVE Database |
Page 1 of 1