llmf

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

• Third-party content exposure detected (high risk: 0.80). This skill's workflow explicitly includes downloading and running models from an external library (see "x llmf model download llava/v1.5-7b/q4_k.gguf" in SKILL.md), which ingests untrusted third-party model files whose outputs can materially influence agent behavior and thus enable indirect prompt injection.