customer-research
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from multiple sources.
- Ingestion points: Untrusted data enters the agent context via customer interview transcripts, survey results, support tickets (Mode 1), and external platforms such as Reddit, G2, and Hacker News (Mode 2).
- Boundary markers: The instructions focus on extracting "exact quotes" but do not define specific technical delimiters or instructions to disregard potential commands embedded within the research material.
- Capability inventory: The skill consists entirely of markdown instructions and does not include any executable scripts, subprocess calls, or network communication capabilities.
- Sanitization: The skill relies on manual qualitative analysis (e.g., confidence level labeling and bias checks) rather than automated technical sanitization of the ingested data.
Audit Metadata