ui-ux-pro-max
Fail
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute commands with elevated privileges using sudo to install Python on Debian-based systems (sudo apt update && sudo apt install python3).
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection where user-supplied keywords are directly interpolated into a shell command: python3 .claude/skills/ui-ux-pro-max/scripts/search.py "".
- [PROMPT_INJECTION]: Mandatory evidence chain: 1. Ingestion points: User requirements collected in Step 1; 2. Boundary markers: Double quotes used in the shell command template; 3. Capability inventory: Execution of system commands and local Python scripts; 4. Sanitization: No evidence of input validation or escaping for the keywords.
- [COMMAND_EXECUTION]: The workflow requires executing multiple shell commands to check environment versions and run script files.
Recommendations
- AI detected serious security threats
Audit Metadata