narrative-snippets

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly tells the agent to read and extract story beats from source material such as "podcast transcripts, interviews, raw notes" (see "Use this skill when transforming podcast transcripts..." and "When processing a podcast or interview transcript, search for:"), which are untrusted/user-generated third‑party inputs that the agent ingests and uses to drive its output and actions, creating a clear opportunity for indirect prompt injection.