mobile-security

SUSPICIOUS. The skill is internally consistent for mobile app pentesting, but it grants an AI agent substantial offensive security capability and relies on numerous third-party tool installs, including some unpinned GitHub/PyPI paths. I found no confirmed malware, no hidden credential theft, no unrelated exfiltration endpoint, and no load-time pre-execution payloads, but the skill remains high-risk because its actual purpose is agent-enabled security testing and bypass activity.