marketing-agency

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow can ingest outsider-authored free text from user-submitted live websites/profiles during “Live URL / profile audit mode” (Step 1) and then use that content as starting context (Step 1.3/Step 1.19-26, plus shared patterns Step 15-17).

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

• Potentially malicious external URL detected (high risk: 0.90). The shared patterns include an install command that fetches and installs code at runtime from https://github.com/vercel-labs/agent-browser (npx skills add https://github.com/vercel-labs/agent-browser --skill agent-browser), which would retrieve and execute remote code that can change agent behavior.