llm-intelligent-public-opinion-analytics

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

• Third-party content exposure detected (high risk: 0.95). The skill’s runtime workflow ingests outsider-authored free text by crawling public platform pages/comments (e.g., via the browser automation in hotsearchcrawler / VideoAnalyzer.extract_content(url=...) with extract_comments=True), then passes the extracted news_content/comment text into the LLM via llm.analyze(content=news_content, ...).

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill includes explicit sudo commands (e.g., sudo mv/chmod to /usr/local/bin), and instructions that modify system-level locations and invoke systemctl, which require elevated privileges and therefore push the agent to change the machine state.