assistant-ui

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly instructs the agent to "Always consult https://assistant-ui.com/llms.txt for latest API", which requires fetching and interpreting content from a public website as part of the workflow and could materially influence runtime/API behavior, creating a vector for indirect prompt injection.