prompt-engineering

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md "Web Research Prompts" (<research_guidelines>) explicitly instructs the agent to "Browse the web" and use web-derived sources/citations, which requires fetching and interpreting open/public third-party (potentially user-generated) content as part of its workflow and can materially influence tool use and decisions.