runpod

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow uses ClawLink to “connect RunPod” and then the agent runs actions/describe/run commands, which necessarily fetches RunPod account/pod/cluster details and action metadata from RunPod’s service (outsider-authored free text via vendor API responses) into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs running "npx @useclawlink/cli" which will fetch and execute remote code from the npm package (see https://www.npmjs.com/package/@useclawlink/cli / registry URL), so a runtime external dependency executes code that the agent relies on.