macro-panel

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow loads each chosen analytics-* skill’s SKILL.md and “matching reference” (per step 2 and the sources manifest), and those references are primarily scraped/fetched from public third-party web pages/books/newsletters (e.g., lynalden.com, hoisington.com, foreign affairs/FT/Project Syndicate, etc.), which are outsider-authored free text that can be ingested into the agent’s LLM context.