write-blog

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

• Third-party content exposure detected (low risk: 0.10). The skill instructs the agent to fetch images and attribution data from Unsplash via API based on user-provided blog topics, which is a low-risk external search/content fetching path.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to execute curl commands against Semrush (api.semrush.com) and Unsplash (api.unsplash.com) using environment variables for runtime data and image retrieval, constituting an operational third-party dependency with weak or unknown user-provisioned publisher provenance.