autowhisper

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow polls AutoWhisper’s API and then relays the returned assistant content into the agent/user context (RESP → jq -r ... .content), which is outsider-authored free text produced by a third-party service (AutoWhisper CMO), not the operating user.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime API calls to the AutoWhisper CMO (e.g. https://autowhisper.xyz/api/cmo/message, https://autowhisper.xyz/api/cmo/messages/$MID, https://autowhisper.xyz/api/cmo/confirm) whose returned "assistant" content directly controls agent instructions and is required for the skill to function.