blackbaud

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

• Third-party content exposure detected (high risk: 0.75). The skill uses ClawLink to perform Blackbaud actions via a browser/OAuth flow, and at runtime the agent will read Blackbaud web UI content (outsider-authored free text from Blackbaud pages) to decide what to do.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires running the remote CLI via npx (npx @useclawlink/cli), which fetches and executes code from the ClawLink npm package (see https://www.npmjs.com/package/@useclawlink/cli) at runtime and therefore can execute remote code and control agent actions.