baoyu-cover-image

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests and deeply analyzes untrusted user-provided reference images (copied to refs/ and passed via --ref) and requires the agent to incorporate "MUST/REQUIRED" elements from those external images into prompts (see references/workflow/reference-images.md and references/workflow/prompt-template.md), so third-party content can materially steer generation behavior.