research-competitors-rival-pass
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from a wide range of untrusted external sources, which creates a surface for indirect prompt injection attacks.
- Ingestion points: The skill retrieves data from competitor websites, news articles, social media platforms (LinkedIn, X), ad libraries (Meta, TikTok, Google), and review sites (Trustpilot, G2, Reddit) as part of its core research workflow.
- Boundary markers: The instructions do not specify any explicit boundary markers or delimiters (such as XML tags or unique string separators) to help the agent distinguish between the skill's instructions and the content fetched from the web.
- Capability inventory: The skill uses web search and page fetching tools. It does not have access to sensitive file-system operations, shell execution, or arbitrary network exfiltration tools, which limits the potential impact of an injection.
- Sanitization: There are no instructions for the agent to sanitize, escape, or validate the content retrieved from external sites before summarizing and reporting on it.
Audit Metadata