ux-audit

Warn

Audited by Socket on Jul 30, 2026

2 alerts found:

SecurityAnomaly
SecurityMEDIUM
scripts/generate-figma-variables.mjs

This CLI utility’s main functionality is safe token parsing/conversion, but it contains a serious security weakness: it uses execSync to run a shell command that embeds a version value originating from an attacker-controlled JSON config (--config). That creates a credible command-injection risk. It also performs runtime download/extraction of an npm tarball without integrity verification, increasing supply-chain attack impact. Treat the tool as unsafe when running with untrusted configs or in sensitive CI environments until execSync usage and download integrity checks are hardened.

Confidence: 82%Severity: 86%
AnomalyLOW
scripts/run-audit-agent.sh

No direct evidence of overt malware (e.g., stealer/backdoor/exfiltration) is present in this Bash wrapper alone. However, it contains high-severity security weaknesses: (1) interactive eval on user-controlled input (command injection risk), (2) automatic GitHub cloning of heuristically matched repos that can introduce untrusted code into the audit workspace, and (3) running the external Claude tool with --dangerously-skip-permissions against the resolved workspace. Additionally, it prints the Figma key and the constructed command line, which can leak sensitive data via logs. Treat this as security-sensitive tooling and remediate by removing eval, validating/escaping inputs, tightening clone behavior (allowlists/integrity checks), and avoiding printing secrets.

Confidence: 78%Severity: 69%
Audit Metadata
Analyzed At
Jul 30, 2026, 12:27 PM
Package URL
pkg:socket/skills-sh/rolemodel%2Frolemodel-design-audit%2Fux-audit%2F@4d7fffc0e93d0004eba65a41942500c967f47f6bc102fed8278ce93dde8bca93
Security Audit — socket — ux-audit