qinghu-bilibili-social

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow fetches and processes public Bilibili video comments, user information, and video descriptions via API tools, ingesting untrusted third-party user text where indirect prompt injections can be introduced.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to make HTTP POST/GET requests to external API endpoints at iqinghu.com during runtime to fetch operational tool lists, search results, video data, and comments, which govern its execution behavior; because iqinghu.com is an arbitrary third-party commercial service domain without strong first-party verification in this context, it qualifies as a weak external operational dependency.