copywriting
Warn
Audited by Snyk on Jun 20, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.85). The workflow’s “Research the Audience” step explicitly pulls language from outsider-authored sources (e.g., G2 reviews, Reddit posts, competitor reviews) at runtime, which can introduce free-form third-party text into the agent’s LLM context via the research ingestion path.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata