replicate-skill

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to inspect model/version schemas and fetch model READMEs via Replicate APIs (e.g., get_models, get_models_readme, search_models in SKILL.md and references), which are user-generated, public content on Replicate that the agent must read and that can materially influence subsequent actions—so it could enable indirect prompt injection.