red-team

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

• Third-party content exposure detected (high risk: 0.70). The skill’s runtime path ingests outsider-authored free text via tools/harness.py reading candidates.jsonl (text field) and passing it to the target/oracle; those candidate texts are not authored by the operating user by definition and can originate from public/other parties’ content.