request-review

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

• Third-party content exposure detected (high risk: 0.75). 该技能在“第一阶段：获取审查结果”允许用户指定外部子代理（claude/cursor/kimi 或用户自定义别名）并通过 CLI 发送提示词；这些代理返回的审查文本会被汇入后续“核实/修复”流程的 LLM 上下文，属于外部方生成的自由文本（OUTSIDER：第三方代理输出）

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). The prompt instructs running external agent CLIs (including flags like --dangerously-skip-permissions and --yolo), accessing the user's terminal config and writing outputs to system temp directories — actions that encourage bypassing permission checks and interacting with the host environment, though it does not explicitly request sudo, modify privileged system files, or create accounts, so the risk is moderate.