agent-browser-core

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime browser operations (e.g., agent-browser open <url> followed by snapshot/get text/attr/value) ingest page body text from an arbitrary URL the operating user supplies, which is outsider-authored free text from public web content into the agent’s LLM context.