coze-image-gen

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires printing generated image URLs verbatim (SKILL.md:34-35) including query parameters such as sign, and also depends on a configured API key (SKILL.md:32), which forces the agent to handle and output secret tokens intact, creating an exfiltration risk.

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

• Third-party content exposure detected (low risk: 0.10). The only runtime text input is the user-supplied --prompt CLI argument (and optional --image <url> reference) that is directly passed to the Coze image generation SDK for processing, but there is no outsider-authored feed/queue/ticket/email/chat content being monitored or ingested by the workflow without explicit user-provided selection.