anti-ai-slop-terminology

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

• Third-party content exposure detected (high risk: 0.75). The required workflow instructs the agent to “browse” and “use evidence” by searching external sources when terminology is disputed/high-impact, which would ingest outsider-authored free text from public web pages into the LLM context at runtime.