ui-ux-pro-max
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill workflow involves searching a local CSV database based on user-provided keywords and then using the retrieved guidelines to implement design code. This presents a surface for indirect injection if the database content were to contain instructional text designed to influence the agent.
- Ingestion points: User-provided search queries processed by
scripts/search.py(referenced inSKILL.md). - Boundary markers: The instructions in
SKILL.mddo not include specific delimiters or warnings to the agent to distinguish between design data and behavioral instructions. - Capability inventory: The skill is authorized to execute its own local Python scripts (
search.py,core.py) and perform file-system reads within its own directory. - Sanitization: The
core.pyscript includes a basic tokenization function that uses regular expressions to remove punctuation and filter input. - [PRIVILEGE_ESCALATION]: The
SKILL.mdfile provides setup instructions that include the use of administrative commands (sudo apt install python3) for Linux environments. While this is a standard procedure for installing prerequisites, it involves the execution of high-privilege commands during the environment preparation phase.
Audit Metadata