create-voice

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

• Third-party content exposure detected (high risk: 1.00). The skill's required COLLECT step and sample guidance (references/sample-collection.md and SKILL.md Step 1) instruct collecting and ingesting user-generated public content (Reddit, Hacker News, forums, social media) and the workflow (Step 2/3) requires the agent to read/interpret those samples to build rules and drive generation/validation, which clearly exposes it to untrusted third‑party content that could enable indirect prompt injection.