marketing-pr

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

• Third-party content exposure detected (medium risk: 0.30). In “Research Mode: Media & PR Intelligence”, the workflow uses agent-browser to open and then “get text body” from third-party pages/feeds like Google News, Connectively sources, LinkedIn search results, and competitor backlink-checker pages—so outsider-authored free text from these sources is ingested without first selecting a specific item.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill's shared-patterns setup instructs installing remote code via the command that fetches and runs https://github.com/vercel-labs/agent-browser (npx skills add ...), which is a required runtime dependency for agent-browser-based research and therefore downloads and executes external code.