byted-kickart-saliency-segmenter

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly asks the agent to prompt the user to paste ACCESS_KEY_ID/SECRET_ACCESS_KEY into chat and then embed those secret values verbatim into export commands, requiring the LLM to handle/output secrets directly.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). High-risk: the SKILL documentation explicitly instructs users to paste their ACCESS_KEY_ID/SECRET_ACCESS_KEY into chat (soliciting credentials), and the code logs request headers (including Authorization) to persistent logs—this combination enables credential capture and exfiltration.

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

• Third-party content exposure detected (medium risk: 0.30). 源文本由外部用户在运行时通过“公网URL/ZIP等图片tos链接”提交，再在必经的图片预处理与下载步骤中被脚本用请求库拉取为图片数据并进一步用于抠图（curl/requests.get 的URL来自用户输入）。