copaw-ai-assistant

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

• Third-party content exposure detected (medium risk: 0.30). The workflow monitors chat platforms (DingTalk, Feishu, Discord, Telegram, etc.) where outsiders can send messages directly into the agent's active channel.

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

• Potentially malicious external URL detected (high risk: 0.70). The installation instructions use shell scripts fetched via curl and powershell from copaw.agentscope.io to set up and execute the software at runtime, presenting a weak/unknown operational dependency.