trends-keywords
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and process external content from user-provided URLs and automated web searches.
- Ingestion points: Data enters the agent context through the 'Inspeção de Contexto' and 'Pesquisa Estratégica' phases described in
SKILL.md, where the model reads external web pages and search trends. - Boundary markers: The instructions do not define specific delimiters (such as XML tags or unique markers) to isolate the retrieved web content from the core instructions.
- Capability inventory: The skill is restricted to web search and report generation. It does not include high-risk capabilities such as file system modification, command-line execution, or unauthorized network operations beyond the search tool.
- Sanitization: No explicit sanitization or pre-processing steps for external content are defined; the skill relies on the underlying model's safety protocols for handling untrusted data.
Audit Metadata