ljg-invest

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

• Third-party content exposure detected (high risk: 1.00). Flagged because the SKILL.md explicitly instructs the agent to fetch public company financials and industry data via a "Research skill or subagent" and to ingest user- or third-party-provided materials (e.g., meeting notes, pitch decks, founder interviews), meaning the agent will read and act on untrusted public/web-sourced content as part of its workflow ("知名公司只给名字即可——用 Research skill 或 subagent 抓取最新财报和行业数据").