derive-client

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

• Third-party content exposure detected (high risk: 0.85). Yes—step 1’s runtime HAR capture ingests outsider-authored free text/HTTP response bodies from the target website (agent-browser loads the site content and embeds response .content.text into the HAR), which then becomes LLM-readable prose via the later jq extraction steps.