minimax-multimodal-toolkit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This skill explicitly tells the agent to ask the user for their MINIMAX_API_KEY and then help configure .env or openclaw.json by embedding that key (e.g., MINIMAX_API_KEY="your-key-here"), which requires the LLM to handle and output secret values verbatim.

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly accepts and ingests arbitrary external media URLs/files as runtime inputs (e.g., references/image-api.md documents subject_reference.image_file can be a "public URL" and scripts/image/generate_image.sh supports --ref-image with http(s) URLs), and video scripts likewise accept first-frame/subject images from URLs, so the agent will fetch and use untrusted, user-provided third‑party content that can materially influence generation behavior.