Replicate Automation

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

• Third-party content exposure detected (low risk: 0.10). The agent reads model schemas and README files from Replicate via active user-specified searches/fetches, which is classified as a low-risk source since the agent must select specific models.

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

• Potentially malicious external URL detected (medium risk: 0.30). The referenced URL https://rube.app/mcp points to an external MCP server setup endpoint, which represents an operational runtime dependency providing agent tools and integration behavior.