aliyun-happyhorse-i2v

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

• Third-party content exposure detected (high risk: 0.70). The skill accepts arbitrary public HTTP/HTTPS "first_frame" image URLs (see "Normalized interface" in SKILL.md and the --first-frame argument in scripts/i2v_happyhorse.py) which are untrusted, user-provided resources that the model ingests to generate video output and thus could carry embedded/indirect instructions that influence model behavior.