hmos-multidevice-scenario-entry

Warn

Audited by Socket on Jul 9, 2026

2 alerts found:

Securityx2
SecurityMEDIUM
scripts/remote_load.sh

This module is a remote installer that fetches untrusted artifacts, extracts them, and deploys them into platform directories. While it does not show overt backdoor/exfiltration behavior, it has substantial supply-chain and host-safety risks: unauthenticated downloads (no integrity/signature), potentially unsafe zip extraction via unzip without explicit traversal protections, and destructive rm -rf operations on cache and target paths derived from user-controlled inputs. A compromised upstream artifact or crafted inputs could lead to file clobbering or installation of malicious/unexpected content into local environments.

Confidence: 72%Severity: 78%
SecurityMEDIUM
SKILL.md

SUSPICIOUS: the skill's routing purpose is coherent, but it extends trust by auto-installing additional scene skills through an unseen remote_load.sh path. The main issue is unverifiable transitive skill loading, not confirmed malicious behavior in the shown file.

Confidence: 84%Severity: 78%
Audit Metadata
Analyzed At
Jul 9, 2026, 08:06 AM
Package URL
pkg:socket/skills-sh/IsKenKenYa%2Fskills%2Fhmos-multidevice-scenario-entry%2F@10cd2b445b9c2f5008fdd627005e59a15619a1a0eb8ec8dafda10e7c255db790
Security Audit — socket — hmos-multidevice-scenario-entry