nano-banana-pro-background-replace

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The README instructs users to pass API keys via a --api-key CLI flag and shows the key prefix (sk-api-*), which encourages embedding secret values verbatim in commands/outputs (an exfiltration risk if the LLM is asked to generate or run such commands).

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

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored text from --prompt (and optional --system) is directly sent as messages/prompt in skill_generate() → cmd_image() → AiHiveClient.generate_image() to the remote AI Hive generation/image endpoint, so free text provided by the caller is ingested at runtime without any filtering or item selection step.