social-content

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow includes monitoring and parsing community-authored Reddit posts and comments via the Reddit API for content ideas and sentiment analysis without prior selection of a specific item.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill instructs the agent to pull documentation from an external URL (https://engagemate.app) to interact with a third-party service during runtime, constituting an unofficial operational dependency.