codex-pet

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

• Third-party content exposure detected (low risk: 0.10). The skill’s required workflow ingests outsider-authored free text only insofar as the user supplies a SOURCE_URL and pet prompt parameters, but the README does not describe any runtime monitoring/consumption of attacker-supplied free text from an external feed/inbox/ticket/queue—only the user’s explicitly provided prompt JSON and a fetched image.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill fetches the source image at runtime (SOURCE_URL="https://.../source.png") which is passed directly into the RunComfy GPT Image 2 edit call and therefore can supply image-based prompt-injection content that controls the model output.