Natural-Language Policy Violations
Medium
- Confidence
- 84% confidence
- Finding
- The skill metadata explicitly constrains output to Chinese-language short-video covers, which can override or conflict with a user's language preference without explicit opt-in. This is not a direct code-execution or data-exfiltration issue, but it can cause unauthorized behavior, reduce user control, and create prompt-steering risk when the agent invokes the skill implicitly.
