browserstack

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime “results” workflow calls BrowserStack MCP tools (e.g., browserstack_get_builds / browserstack_get_sessions / browserstack_get_session / browserstack_get_logs) to ingest build/session/log text from BrowserStack’s external service into the agent’s LLM context, which is outsider-authored free text.