influencer-brief
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied campaign details (brand, product, and goal) and interpolates them into a structured brief. This ingestion of untrusted data creates a potential surface for indirect prompt injection attacks.\n
- Ingestion points: The skill ingests external content via the
$ARGUMENTSvariable inSKILL.mdor via natural language responses when prompted for brand details.\n - Boundary markers: The instructions do not define clear delimiters (e.g., XML tags or triple quotes) or provide 'ignore embedded instructions' warnings for the agent when processing the user input.\n
- Capability inventory: The skill operates with access to the
ReadandWritetools, enabling it to load benchmark data and potentially save documents.\n - Sanitization: There are no explicit validation steps or sanitization filters applied to the campaign seed data before it is integrated into the final formatted document.
Audit Metadata