wandb-autoresearch

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow reads W&B content controlled by the provider user (outsider) by sampling the project’s recent runs and their logged/used artifacts (e.g., job/code artifacts) and listing queues to inspect capacity/state, all via wandb.Api().runs(...), run.logged_artifacts(), and run.used_artifacts() in the Understand stage before running experiments.