newsclawd

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

• Third-party content exposure detected (high risk: 0.80). The skill fetches and interprets public, user-generated content — e.g., scripts/rss.py uses urlopen(url) to pull arbitrary RSS/Atom feeds and SKILL.md explicitly describes monitoring RSS feeds, subreddits, web scraping and X/Twitter accounts — and those feed contents are used to trigger alerts and actions, so untrusted third-party content can materially influence behavior.