ai-content-pipeline

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly includes a "Research -> Content -> Distribution" pattern and lists the tavily/search-assistant in Pipeline Building Blocks and a "Blog to Video Pipeline" step that summarizes "" (and accepts arbitrary URLs like "https://your-music.mp3"), indicating the skill ingests public/untrusted web content which the agent is expected to read and which can change subsequent pipeline actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The Quick Start contains the command "curl -fsSL https://cli.inference.sh | sh", which fetches and executes a remote install script at runtime (https://cli.inference.sh), making that URL a required external dependency that runs remote code.