hubspot

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow uses ClawLink to access HubSpot via the user’s OAuth session and then returns HubSpot CRM data (e.g., contact/company/deal/ticket fields) as readable text into the agent context; this is outsider-authored content in the sense that it originates from a third-party SaaS system rather than the operating user’s own authored prompts/files.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs the agent to run npx @useclawlink/cli commands which fetch and execute remote code from the ClawLink/npm package (see https://www.npmjs.com/package/@useclawlink/cli and https://claw-link.dev), so external content is fetched at runtime and controls behavior.