runpod-vllm

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

• Third-party content exposure detected (low risk: 0.10). The required workflow reads the user-selected Hugging Face model card and config.json (SKILL.md Step 1) and the public vLLM recipes repo, which are outsider-published content the agent fetches by choosing a specific source, so the attacker can publish poison but cannot force delivery.