content-strategy

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md "Customer Research for Content" workflow explicitly instructs mining untrusted public sources (e.g., "Community forums (Reddit, Slack, Discord)", product reviews, competitor blogs) and the content_brief_generator.py output includes "Analyze top 5 ranking pages for ''", which require ingesting and acting on arbitrary public web/user-generated content that could influence agent decisions.