us-market-bubble-detector

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

• Third-party content exposure detected (high risk: 1.00). The skill's required Phase 1 data-collection and Implementation Guide explicitly instruct the agent to web_search and ingest public third-party sources (e.g., CBOE/Yahoo Finance/FINRA/Barchart/Renaissance Capital, Google Trends, Twitter/Reddit/news APIs as listed in SKILL.md and references/implementation_guide.md), and those untrusted results are used to compute scores and drive recommendations, which clearly allows third-party content to materially influence agent actions.