nano-banana-pro-multi-reference-image

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill documentation explicitly allows passing API keys via a --api-key command-line flag (and references the sk-api-* format), which encourages embedding secrets directly in commands/outputs and therefore creates an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored free text entered in the required workflow as the CLI --prompt (and optional --system) is sent verbatim in the messages body to POST /openapi/v1/chat/text or prompt to POST /openapi/v1/generation/image/generation/video, so the agent runtime ingests untrusted text before any item selection.