tool-ui-ux-pro-max
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates entirely locally using bundled data files and a standard Python search implementation. All functionalities are limited to retrieving information from the provided CSV files.
- [COMMAND_EXECUTION]: The agent uses the
python3command to run a search script included in the skill's own directory. This script is used solely for querying the internal knowledge base and does not involve any dangerous system commands or privilege escalation. - [DATA_EXFILTRATION]: No evidence of data exfiltration was found. File access is strictly limited to the skill's own data files through hardcoded configuration, and the skill does not import or use any libraries for network operations.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user queries to retrieve design tokens from internal CSVs, the query processing is done via safe tokenization and ranking. The retrieved content consists of informational design guidance, which poses no risk of overriding agent behavior.
Audit Metadata