user-research

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to its data-processing workflow.\n
  • Ingestion points: The skill ingests untrusted external data using WebFetch, WebSearch, and Read tools to analyze user interviews and competitor websites.\n
  • Boundary markers: The instructions lack explicit delimiters or 'ignore' directives to prevent the model from inadvertently executing commands that might be embedded in research transcripts or fetched web content.\n
  • Capability inventory: The presence of Write, Edit, and WebFetch tools creates a potential risk where malicious input could influence file system changes or further network requests.\n
  • Sanitization: No input validation or sanitization logic is defined for the content processed during synthesis steps.\n- [EXTERNAL_DOWNLOADS]: The skill utilizes the WebFetch and WebSearch tools to retrieve data from external web sources. While this is necessary for its UX research functionality, it involves network interactions with non-whitelisted domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 02:04 PM