parlor

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

• Third-party content exposure detected (medium risk: 0.30). The agent skill monitors and consumes messages from shared parlor rooms via parlor read and parlor wait, which contain free text posted by arbitrary external participants.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs downloading and executing a remote bash script from parlor.sh (https://parlor.sh/cli), which constitutes an operational dependency with unknown/third-party provenance.