ui-ux-pro-max
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill ingests untrusted data from search results and formats it for the agent's context. • Ingestion points: User-supplied query and search results from core and design_system modules processed in scripts/search.py. • Boundary markers: The format_output function uses basic markdown headers (e.g., ### Result) which provide minimal isolation between data and instructions. • Capability inventory: The script is primarily a data retrieval and formatting utility; no direct file modification or network exfiltration capabilities are present in the analyzed file. • Sanitization: There is no evidence of content validation or instruction-filtering for the retrieved data before it is presented to the agent.
Audit Metadata