finance-sentiment

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

• Third-party content exposure detected (high risk: 0.90). This skill's SKILL.md and references/api_reference.md explicitly instruct the agent to call Adanos API endpoints (e.g., /reddit/stocks/v1/compare, /x/stocks/v1/compare, /news/... and /polymarket/...) to retrieve and synthesize signals (and optionally detail fields like top mentions) derived from public, user-generated sources (Reddit, X, news, Polymarket), so the agent ingests untrusted third-party content that can materially influence its outputs.