phantombuster

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

• Third-party content exposure detected (medium risk: 0.30). The skill reads CRM, lead, and agent output data from PhantomBuster, which can contain outsider-authored free text such as scraped profile data or messages.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs execution of npm packages via npx referencing remote URLs and services without verified first-party provenance, constituting a weak operational dependency.