social-intelligence

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md explicitly instructs calling the external endpoints (e.g., https://stableenrich.dev/api/reddit/search and https://stableenrich.dev/api/reddit/post-comments) to fetch Reddit posts and comments (user-generated content) and then parse/summarize those results in workflows like "Brand Monitoring" and "Community Sentiment," so untrusted third‑party content can influence agent decisions.