fal-ai-media

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

• Third-party content exposure detected (high risk: 0.70). Yes — the SKILL.md explicitly shows the MCP consuming external media URLs (e.g., the ThinkSound example's "video_url" and the image-to-video "image_url" fields) which the agent/MCP is expected to fetch and process, so untrusted public content could materially influence generation behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires running "npx -y fal-ai-mcp-server" as the MCP command, which fetches and executes remote code from the npm registry (e.g., https://www.npmjs.com/package/fal-ai-mcp-server) at runtime, making it a required external dependency that executes remote code.