bibi-feed

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

• Third-party content exposure detected (medium risk: 0.30). The workflow reads a video/podcast feed (bibi feed) from user-subscribed channels where titles and metadata originate from third-party video platforms, allowing channel creators to inject untrusted free text into the agent's feed.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes a Linux installation command that pipes a remote shell script from https://bibigpt.co/install.sh directly into bash at runtime, which executes remote code.