Social Media Manager

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md and references/advanced-patterns.md explicitly direct the agent to use public web search and ingest/analyze user-generated social content (e.g., competitor Twitter/X profiles, Reddit posts, and external newsletters/blog links) as part of workflows that drive drafting, scheduling, and strategy decisions, so untrusted third-party content can materially influence actions.