runcomfy-cli

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

• Third-party content exposure detected (low risk: 0.10). The workflow executes RunComfy CLI commands which invoke cloud AI models; while model inputs and outputs (like web search or image assets) can theoretically contain indirect injection vectors, the primary runtime is a command-line interface controlled by the user/agent rather than a monitored outsider submission feed.

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

• Potentially malicious external URL detected (medium risk: 0.30). The installation instructions include an external command to install a skill via npx skills add agentspace-so/runcomfy-agent-skills, which is an inbound operational dependency from an unofficial third-party repository.