wandb

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

• Third-party content exposure detected (high risk: 0.90). The skill's scripts (e.g., scripts/characterize_run.py, compare_runs.py, run_details.py, watch_runs.py) call wandb.Api() and directly fetch run data (api.run(...), run.scan_history()/run.history(), run.config, run.summary, run.tags, etc.), thereby ingesting untrusted, user-generated W&B project/run content which the agent parses and uses to produce health checks, verdicts, and follow-up outputs.