ui-ux-pro-max
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCOMMAND_EXECUTION
Full Analysis
- [Privilege Escalation] (HIGH): The prerequisites section in
SKILL.mdinstructs the user to executesudo apt update && sudo apt install python3. Requesting administrative privileges is a high-risk operation that should be avoided in agent skills. - [Indirect Prompt Injection] (HIGH): The workflow in
SKILL.md(Step 2) usespython3 .claude/skills/ui-ux-pro-max/scripts/search.py "<keyword>"to process user input. This pattern is vulnerable to command injection if the<keyword>extracted from the user request contains shell metacharacters. - Ingestion points: User requirements extracted in Step 1 (Product type, Style keywords, etc.).
- Boundary markers: None. Input is placed directly inside double quotes in a shell command.
- Capability inventory: Shell command execution via
python3subprocess calls. - Sanitization: No evidence of sanitization or escaping of the
<keyword>parameter, allowing an attacker to execute arbitrary commands by providing a keyword like"; touch /tmp/pwned; ".
Recommendations
- AI detected serious security threats
Audit Metadata