blog-post-writer

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to fetch live content from public sites — e.g., https://nicknisi.com/posts and the Wikipedia API page for "Wikipedia:Signs_of_AI_writing" — and to parse those pages to calibrate voice and run the AI-slop checks, meaning untrusted third‑party content is read and can materially influence drafting and tool decisions.