ui-ux-pro-max
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates as a localized information retrieval system. It utilizes standard Python libraries (argparse, csv, math, re) to perform BM25 search operations on provided CSV datasets.
- [COMMAND_EXECUTION]: The SKILL.md documentation includes standard environment setup instructions for users (e.g., brew, apt-get, winget). These are instructional for prerequisites and do not involve the agent executing privileged commands automatically or without oversight.
- [DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive environment configurations were detected across the 21 files analyzed.
- [PROMPT_INJECTION]: The instructions focus on UI/UX best practices and do not contain attempts to bypass agent safety filters or override system instructions.
- [REMOTE_CODE_EXECUTION]: The Python scripts (core.py and search.py) are self-contained and do not download or execute remote code. All data is processed from local CSV files.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user queries through its search script, the output is restricted to matching records from the curated CSV datasets, effectively mitigating the risk of untrusted data influencing the agent's core behavior.
Audit Metadata