game-character-sprites

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflow explicitly requires grounding generation on user-provided reference images (SKILL.md Agent Workflow step 2: "Ground the character from the user reference" / "use the attached image from chat context") and the manifest validator permits external "image_url" references (scripts/validate_run_manifest.py ALLOWED_REFERENCE_TYPES), so the agent ingests arbitrary/untrusted third-party image content which directly informs generation, validation, and regeneration decisions.