research-topic

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 Step 2 explicitly instructs the agent to use tool web_search and tool web_research to run multiple searches against public web sources and synthesize those findings into a report, so it fetches and reads untrusted third‑party content (open web pages, papers, news) that can materially influence its outputs.