xpoz-social-search

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md and Tool Reference explicitly show the skill queries public social media (Twitter, Instagram, Reddit) using tools like getTwitterPostsByKeywords / getInstagramPostsByKeywords / getRedditPostsByKeywords, so the agent ingests untrusted, user-generated posts as part of its workflow which could contain instructions that influence subsequent actions.