humanize

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

• Third-party content exposure detected (high risk: 0.70). The skill ingests arbitrary user-provided text via the /humanize command and sends/returns that untrusted user-generated content through the HumanizerAI API (https://humanizerai.com), meaning the agent reads and displays third-party content that could carry indirect prompt injections.