kelly-content

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow can ingest outsider-authored free text via the user-supplied “source” (e.g., blog post/transcript/outline/product announcement) that is passed into the agent for extraction and generation, which then becomes LLM context; this is not limited to trusted/bundled material.