engagement-tracker

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly scrapes and ingests user-generated content from public sites (Reddit via "Step 2: Scrape Reddit Comment Metrics" and Twitter/X via "Step 3: Scrape Twitter Reply Metrics" in SKILL.md), reads comment/tweet text and metrics as part of its workflow, and uses that content to decide actions (e.g., flagging replies for response and driving weekly recommendations), so untrusted third-party content could influence behavior and enable indirect prompt injection.