review-model-performance

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow ingests outsider-authored free text via the tile’s evals/*/task.md and scenario content (e.g., tessl eval view <id> results and scenario definitions), which are not guaranteed to be authored by the operating user and are read/used to drive the evals.