community-engagement

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly directs searching and reading public, user-generated content from Reddit, Facebook groups, and Discord (see "Step 1: Define Search Queries", "Step 2: Surface Relevant Conversations", and the "WebFetch" integration), which the agent is expected to ingest and use to craft summaries and engagement actions.